In this paper, we present a literature survey about drowsy driving detection using PERCLOS metric that determines the percentage of eye closure. This metric determines that an eye is closed if the percentage of eye closure is 80% or above. When this percentage is observed for multiple frames of a video camera feed, the driver is determined to be in an unsafe fatigue status. In our research, we found that the PERCLOS metric had a 0.79 to 0.87 correlation coefficient value which exceeds the 0.7 R value needed to be considered a strong correlation coefficient. A higher value than 0.7 indicates a more linear relationship which means that the metric is dependable [1].
How to Implement the Digital Medicine in the FutureYoon Sup Choi
Digital Pathologist uses image analysis to extract quantitative features from digitized pathology slides to predict cancer survival. It segments images into epithelial and stromal regions and measures thousands of morphological features. These include standard metrics like nuclear size as well as higher-level relationships between image objects. Models are trained on annotated slides from cancer patients with known survival outcomes. The system was able to build a prognostic model from one dataset and validate it on a separate, independent dataset, identifying novel image-based features associated with survival in the process. This automated digital pathology approach could help standardize quantitative analysis and uncover new biological insights compared to traditional visual examination alone.
Computer vision is a field of artificial intelligence that trains computers to interpret and understand the visual world. Using digital images from cameras and videos and deep learning models, machines can accurately identify and classify objects — and then react to what they “see.”
IRJET - Deep Multiple Instance Learning for Automatic Detection of Diabetic R...IRJET Journal
This document describes a proposed method for using deep multiple instance learning to automatically detect diabetic retinopathy in retinal images. Diabetic retinopathy is a complication of diabetes that can cause vision loss or blindness. The proposed method treats retinal images as "bags" containing "instances" of image patches. A deep learning model is trained using only image-level labels to both detect diabetic retinopathy images and identify lesions within images. The model first preprocesses images to normalize factors like scale and illumination. It then segments lesions and extracts features before classifying images using convolutional neural networks. The goal is to provide explicit locations of lesions to aid clinicians while leveraging large datasets typically required for deep learning.
Comparative Study of Classification Method on Customer Candidate Data to Pred...IJECEIAES
Leasing vehicles are a company engaged in the field of vehicle loans. Purchase by way of credit becomes a mainstay because it can attract potential customers to generate more profit. But if there is a mistake in approving a customer candidate, the risk of stalled credit payments can happen. To minimize the risk, it can be applied the certain data mining technique to predict the future behavior of the customers. In this study, it is explored in some data mining techniques such as C4.5 and Naive Bayes for this purpose. The customer attributes used in this study are: salary, age, marital status, other installments and worthiness. The experiments are performed by using the Weka software. Based on evaluation criteria, i.e. accuracy, C4.5 algorithm outperforms compared to Naive Bayes. The percentage split experiment scenarios provide the precision value of 89.16% and the accuracy value of 83.33% wheres the cross validation experiment scenarios give the higher accuracy values of all used k-fold. The C4.5 experiment results also confirm that the most influential instant data attribute in this research is the salary.
IRJET- Effect of Principal Component Analysis in Lung Cancer Detection us...IRJET Journal
This document discusses using machine learning techniques to detect lung cancer from data more accurately and quickly. It summarizes that lung cancer is a leading cause of cancer death worldwide. Current diagnosis methods like CT scans can detect small lung lesions but take time. The document proposes using machine learning algorithms on lung cancer data to classify and detect cancer, aiming to diagnose it earlier. It discusses collecting lung cancer data from a repository and filtering/classifying it using methods like J48, principal component analysis, and comparing results to find the best detection method.
IRJET- Survey on Face Detection MethodsIRJET Journal
The document reviews 15 papers on various face detection methods published between 2013 and 2018. It finds that the most popular feature extraction method is skin color segmentation, which achieves detection rates of 88-98%. The Viola-Jones method typically detects face regions as well as other body parts at a rate of 80-90%. Common face detection methods reviewed include skin color segmentation, Viola-Jones, Haar features, 3D mean shift, and Cascaded Head and Shoulder Detection. OpenCV, Python or MATLAB are typically used to implement real-time face detection systems.
Evaluation of Iris Recognition System on Multiple Feature Extraction Algorith...Editor IJCATR
Multi-algorithmic approach to enhancing the accuracy of iris recognition system is proposed and
investigated. In this system, features are extracted from the iris using various feature extraction algorithms,
namely LPQ, LBP, Gabor Filter, Haar, Db8 and Db16. Based on the experimental results, it is demonstrated
that Mutli-algorithms Iris Recognition System is performing better than the unimodal system. The accuracy
improvement offered by the proposed approach also showed that using more than two feature extraction
algorithms in extracting the iris system might decrease the system performance. This is due to redundant
features. The paper presents a detailed description of the experiments and provides an analysis of the
performance of the proposed method.
Evaluation of Iris Recognition System on Multiple Feature Extraction Algorith...Editor IJCATR
Multi-algorithmic approach to enhancing the accuracy of iris recognition system is proposed and investigated. In this system, features are extracted from the iris using various feature extraction algorithms, namely LPQ, LBP, Gabor Filter, Haar, Db8 and Db16. Based on the experimental results, it is demonstrated that Mutli-algorithms Iris Recognition System is performing better than the unimodal system. The accuracy improvement offered by the proposed approach also showed that using more than two feature extraction algorithms in extracting the iris system might decrease the system performance. This is due to redundant features. The paper presents a detailed description of the experiments and provides an analysis of the performance of the proposed method.
How to Implement the Digital Medicine in the FutureYoon Sup Choi
Digital Pathologist uses image analysis to extract quantitative features from digitized pathology slides to predict cancer survival. It segments images into epithelial and stromal regions and measures thousands of morphological features. These include standard metrics like nuclear size as well as higher-level relationships between image objects. Models are trained on annotated slides from cancer patients with known survival outcomes. The system was able to build a prognostic model from one dataset and validate it on a separate, independent dataset, identifying novel image-based features associated with survival in the process. This automated digital pathology approach could help standardize quantitative analysis and uncover new biological insights compared to traditional visual examination alone.
Computer vision is a field of artificial intelligence that trains computers to interpret and understand the visual world. Using digital images from cameras and videos and deep learning models, machines can accurately identify and classify objects — and then react to what they “see.”
IRJET - Deep Multiple Instance Learning for Automatic Detection of Diabetic R...IRJET Journal
This document describes a proposed method for using deep multiple instance learning to automatically detect diabetic retinopathy in retinal images. Diabetic retinopathy is a complication of diabetes that can cause vision loss or blindness. The proposed method treats retinal images as "bags" containing "instances" of image patches. A deep learning model is trained using only image-level labels to both detect diabetic retinopathy images and identify lesions within images. The model first preprocesses images to normalize factors like scale and illumination. It then segments lesions and extracts features before classifying images using convolutional neural networks. The goal is to provide explicit locations of lesions to aid clinicians while leveraging large datasets typically required for deep learning.
Comparative Study of Classification Method on Customer Candidate Data to Pred...IJECEIAES
Leasing vehicles are a company engaged in the field of vehicle loans. Purchase by way of credit becomes a mainstay because it can attract potential customers to generate more profit. But if there is a mistake in approving a customer candidate, the risk of stalled credit payments can happen. To minimize the risk, it can be applied the certain data mining technique to predict the future behavior of the customers. In this study, it is explored in some data mining techniques such as C4.5 and Naive Bayes for this purpose. The customer attributes used in this study are: salary, age, marital status, other installments and worthiness. The experiments are performed by using the Weka software. Based on evaluation criteria, i.e. accuracy, C4.5 algorithm outperforms compared to Naive Bayes. The percentage split experiment scenarios provide the precision value of 89.16% and the accuracy value of 83.33% wheres the cross validation experiment scenarios give the higher accuracy values of all used k-fold. The C4.5 experiment results also confirm that the most influential instant data attribute in this research is the salary.
IRJET- Effect of Principal Component Analysis in Lung Cancer Detection us...IRJET Journal
This document discusses using machine learning techniques to detect lung cancer from data more accurately and quickly. It summarizes that lung cancer is a leading cause of cancer death worldwide. Current diagnosis methods like CT scans can detect small lung lesions but take time. The document proposes using machine learning algorithms on lung cancer data to classify and detect cancer, aiming to diagnose it earlier. It discusses collecting lung cancer data from a repository and filtering/classifying it using methods like J48, principal component analysis, and comparing results to find the best detection method.
IRJET- Survey on Face Detection MethodsIRJET Journal
The document reviews 15 papers on various face detection methods published between 2013 and 2018. It finds that the most popular feature extraction method is skin color segmentation, which achieves detection rates of 88-98%. The Viola-Jones method typically detects face regions as well as other body parts at a rate of 80-90%. Common face detection methods reviewed include skin color segmentation, Viola-Jones, Haar features, 3D mean shift, and Cascaded Head and Shoulder Detection. OpenCV, Python or MATLAB are typically used to implement real-time face detection systems.
Evaluation of Iris Recognition System on Multiple Feature Extraction Algorith...Editor IJCATR
Multi-algorithmic approach to enhancing the accuracy of iris recognition system is proposed and
investigated. In this system, features are extracted from the iris using various feature extraction algorithms,
namely LPQ, LBP, Gabor Filter, Haar, Db8 and Db16. Based on the experimental results, it is demonstrated
that Mutli-algorithms Iris Recognition System is performing better than the unimodal system. The accuracy
improvement offered by the proposed approach also showed that using more than two feature extraction
algorithms in extracting the iris system might decrease the system performance. This is due to redundant
features. The paper presents a detailed description of the experiments and provides an analysis of the
performance of the proposed method.
Evaluation of Iris Recognition System on Multiple Feature Extraction Algorith...Editor IJCATR
Multi-algorithmic approach to enhancing the accuracy of iris recognition system is proposed and investigated. In this system, features are extracted from the iris using various feature extraction algorithms, namely LPQ, LBP, Gabor Filter, Haar, Db8 and Db16. Based on the experimental results, it is demonstrated that Mutli-algorithms Iris Recognition System is performing better than the unimodal system. The accuracy improvement offered by the proposed approach also showed that using more than two feature extraction algorithms in extracting the iris system might decrease the system performance. This is due to redundant features. The paper presents a detailed description of the experiments and provides an analysis of the performance of the proposed method.
This is a summary of Siegel's Predictive Analytics. The presentation encourages the reader to actively look at the capabilities of Predictive Analytics as a tool to make a forecast for one entity.
This document summarizes the results of a survey of 81 refractive surgeons on the business impacts of adopting the Intralase femtosecond laser for LASIK flap creation. Key findings include:
1) Practices typically increased LASIK fees by an average of $394 after adopting Intralase to offset costs. Higher fee practices saw smaller increases.
2) Over 80% of patients chose Intralase when given the option, leading more surgeons to make it standard.
3) Practices saw an average 41% increase in LASIK revenue and a 40% decrease in retreatments after adopting Intralase.
This document summarizes medical imaging technologies developed by the Institute for Infocomm Research for screening major ocular diseases in primary care. It describes technologies for automatically grading cataracts, calculating cup-to-disc ratios for glaucoma screening, and detecting various eye conditions like drusen, pathological myopia, and cornea diseases from different image modalities like lens, fundus, OCT, and RETCAM images. It also details the AGLAIA system for automatic glaucoma diagnosis and its validation in a study of 1,676 subjects where it achieved diagnostic performance comparable to expert grading of fundus photos and eye pressure measurements.
IRJET- GDPS - General Disease Prediction SystemIRJET Journal
The document describes a General Disease Prediction System (GDPS) that uses machine learning and data mining techniques to predict diseases based on patient symptoms.
The GDPS first collects patient data, preprocesses it, and extracts relevant features. It then implements the ID3 decision tree algorithm to generate a predictive model and classify diseases. As an admin, one can train the model using sample data. As a user, one can enter symptoms and the trained model will predict the likely disease and recommend precautions.
The GDPS was tested on a dataset of 120 patients and achieved 86.67% accuracy in disease prediction. The system currently covers common diseases but future work involves expanding it to predict more serious or fatal diseases like various cancers
Credit card fraud detection through machine learningdataalcott
This document discusses using machine learning algorithms for credit card fraud detection. It proposes using principal component analysis for feature selection followed by logistic regression and decision tree models. It finds that logistic regression has higher accuracy at 79.91% compared to 71.41% for decision tree. The proposed approach aims to better handle imbalanced data and reduce fraudulent transactions. Future work could implement the approach in Python and produce experimental results.
LAN Based HF Radio Simulator An Approach to Develop an Early Prototype for Lo...YogeshIJTSRD
This document summarizes the key aspects of developing a LAN-based HF radio simulator. It conducted surveys and interviews with users to understand requirements. The surveys found that a simulator could facilitate training by allowing practice without actual radios. A web-based platform was preferred. The document outlines the methodology, which includes specifying user and software requirements. It defines functional and non-functional requirements for the simulator. The simulator aims to allow trainees to practice operating radios in different scenarios while preserving performance records.
A New Proposal for Smartphone Based Drowsiness Detection and Warning System f...YogeshIJTSRD
This paper proposes a Smartphone based system for the detection of drowsiness in automotive drivers. The proposed system uses three stage drowsiness detection technique. The first stage uses the percentage of eyelid closure PERCLOS which is obtained by capturing images with the front camera of the Smartphone with a modified eye state classification method. The system uses near infrared lighting for illuminating the face of the driver during night driving. The second step uses the voiced to the unvoiced ratio VUR obtained from the speech data from the microphone, in the event PERCLOS crosses the threshold. The VUR is also compared with a threshold and if it is a value greater than that of the threshold, it moves on to the next verification stage. In the final verification stage, touch response is required within the stipulated time to declare whether the driver is drowsy or not and subsequently sound an alarm. To awake the driver, a vibrating mechanism is done and also the live GPS location is also sent to an emergency contact. We have studied eight other reference papers for the literature review. The system has three advantages over existing drowsiness detection systems. First, the three stage verification process makes the system more reliable. The second advantage is its implementation on an Android smart phone, which is readily available to most drivers or cab owners as compared to other general purpose embedded platforms. The third advantage is the use of SMS service to inform the control room as well as the passenger regarding the loss of attention of the driver. Abishek K Biju | Godwin Jolly | Asif Mohammed C A | Dr. Paul P Mathai | Derek Joseph "A New Proposal for Smartphone-Based Drowsiness Detection and Warning System for Automotive Drivers" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-5 | Issue-5 , August 2021, URL: https://www.ijtsrd.com/papers/ijtsrd45083.pdf Paper URL: https://www.ijtsrd.com/engineering/computer-engineering/45083/a-new-proposal-for-smartphonebased-drowsiness-detection-and-warning-system-for-automotive-drivers/abishek-k-biju
This document summarizes research on driver drowsiness detection systems. It begins with an abstract describing how convolutional neural networks were used to create a model to classify eyes as open or closed using webcam photos. It then reviews six related works that used techniques like LSTM-CNN architectures, support vector machines, and multi-modal information to detect drowsiness. The document outlines the objectives, methodology, application requirements, and conclusions of the research, which involved using image processing and deep learning models to accurately detect drowsiness in a non-invasive manner to improve road safety.
An efficient system for real time fatigue detectionAlexander Decker
This document summarizes a research paper that proposes an efficient system for real-time fatigue detection. The system uses computer vision and image processing techniques to measure eye closure count, blinking rate, and yawning to detect user fatigue. Face detection is performed using the Viola-Jones algorithm. Abnormalities in eye and mouth behavior are then analyzed to determine if the user is fatigued. The system aims to detect fatigue early enough to avoid accidents in applications where user attentiveness is critical. It is designed to have low time and space complexity, be low cost, and not significantly impact normal user interactions. The proposed approach and algorithm are described, and example results of fatigue detection are provided.
A presentation of Driver drowsiness alert system which can identify whether the driver is attentive or sleepy while driving and hence alert them by a beep when the driver is sleepy.Python and open CV are main technologies used here along with hass cascade algorithm for the same.
Driving without license is the major cause for the road accident and the equivalent monetary losses. This paper is based on virtual reality based driving system which would enhance road safety and vehicle security. This paper helps to limit the vehicle operation on the basics of two parameters-Learn the driving by our own, category (car or bike) of the vehicle for which the driving license is issued. The hardware and software system required to improve our safety and security is developed. This driving system is apt for getting the license without bribe by gathering eye-gaze, Electroencephalography and peripheral physiological data.
When driving long distances, drivers who do not take frequent rests are more likely to get sleepy, a condition that experts say they often fail to identify early enough. Based on eye condition, this research proposes a system for detecting driver sleepiness in real time. A camera is often used to take a sequence of images by the system. In our system, these capture images may be saved as individual frames. The resulting frame is sent into facial recognition software as an input. The image's needed feature (eye) is then extracted. The method creates a condition for each eye and suggests a certain number of frames with the same condition that can be registered.
Driver's drowsiness is the main reason for vehicular accidents. Drowsy driving is the form of impaired driving
that continuously affects a person's ability to drive safely. Continuous restless driving for longer time may result in
drowsiness and cause accidents. In this study, a collaborative system is build which assist the user and identifies
his/her state while driving in order to improve safety by preventing accidents. Based on grayscale image processing,
the position of the driver's face and his/her head movement is analysed. The driver's state identification also includes
the detection of alcohol consumption with the help of sensors.
Enhancing Time Series Anomaly Detection: A Hybrid Model Fusion ApproachIJCI JOURNAL
Exploring and Identifying Anomalies in time-series data is very crucial in today’s world revolve around data. These data are being used to make important decisions; hence, an efficient and reliable anomaly detection system should be involved in this process to ensure that the best decisions are being made. The paper explores other types of anomalies and proposes efficient detection methods which can be used. Anomalies are patterns that deviate from usual expected behavior. These can come from system failures or unexpected activity. This research paper explores the vulnerabilities of commonly used anomaly detection algorithms such as the Z-Score and static threshold approach. Each method used in this paper has its unique capabilities and limitations. These methods range from using statistical methods and machine learning approaches to detecting anomalies in a time-series dataset. Furthermore, this paper explores other open-source libraries that can be used to detect anomalies, such as Greykite and Prophet Python library. This paper serves as a good source for anyone new to anomaly detection and willing to explore.
Driver Drowsiness is a grave issue resulting in many road accidents each year. To evaluate the exact number of sleep related accidents because of the difficulties in detecting whether fatigue was a factor and in assessing the level of fatigue is not currently possible. In this paper the camera will be placed besides the rare view mirror of car in way such that it is in clear view of the frontal face of the driver. This camera will continuously capture the video of driver’s frontal face while driving. The system will detect the frontal face in the image and later the eyes. Depending upon the conditions the system will generate an alert. The focus will be on the system that will accurately monitor the open or closed state of the driver’s eyes in real-time. By monitoring the eyes, it is believed that the symptoms of driver fatigue can be detected early to avoid accidents.
There are different types of algorithms used in eye tracking technologies. These algorithms are divided into two main categories: feature-based and model-based. Feature-based technologies consist of threshold values, which are used to decide the presence or absence of features or determinant factors. While the model-based approach is an iterative search of a model parameter, which is the best fitting model that is a closest match to the image. However, these approaches have significant problems regarding computational speed and accuracy. Similarly, there are different types of eye – tracking technologies, which depend on different types of technologies such as infrared video cameras and other technologies, which require specific calibration and setup and are quite expensive. Therefore, in this paper, we propose an alternative eye–tracking technology using a new eye-tracking algorithm, which is highly portable and independent of any hardware or software systems. In an evaluation the algorithm worked accurately for users with strong dyslexia. Participants had various positive and negative opinions regarding such an auto-detection system. Furthermore, we propose that such technology could be used to automatically modify the content of online material to better suit dyslexic users.
AN ALGORITHM FOR AUTOMATICALLY DETECTING DYSLEXIA ON THE FLYijcsit
There are different types of algorithms used in eye tracking technologies. These algorithms are divided into two main categories: feature-based and model-based. Feature-based technologies consist of threshold values, which are used to decide the presence or absence of features or determinant factors. While the model-based approach is an iterative search of a model parameter, which is the best fitting model that is a
closest match to the image. However, these approaches have significant problems regarding computational speed and accuracy. Similarly, there are different types of eye – tracking technologies, which depend on different types of technologies such as infrared video cameras and other technologies, which require specific calibration and setup and are quite expensive. Therefore, in this paper, we propose an alternative eye–tracking technology using a new eye-tracking algorithm, which is highly portable and independent of any hardware or software systems. In an evaluation the algorithm worked accurately for users with strong dyslexia. Participants had various positive and negative opinions regarding such an auto-detection system. Furthermore, we propose that such technology could be used to automatically modify the content of online material to better suit dyslexic users.
Drowsiness and Alcohol Detection for Accident Prevention using Machine Learningijtsrd
A Drowsy Driver Detection System has been created, utilizing a non meddling machine Vision based absolutely ideas. The framework utilizes a touch monochrome surveillance camera that focuses Directly towards the drivers face and screens the drivers eyes along these lines on watch weakness. In Such a case once exhaustion is identified, an alarm is given to caution the main impetus. This Report depicts the gratitude to see the eyes, and together the gratitude to check if the eyes zone unit open or Closed. The algorithmic standard created is restrictive to any directly unconcealed papers, that was a Primary goal of the venture. The framework manages exploitation data acquired for the Binary form of the picture to go glancing out the edges of the face, that limits the domain of where the Eyes may exist. When the face region is discovered, the eyes zone unit found by registering the flat Averages at stretches the region. Taking into thought the data that eye locales at stretches the face blessing decent power changes, the eyes zone unit put by finding the various force changes at spans the face. When the eyes zone unit set, live the separations between the force changes at spans the consideration zone confirm whether or not the eyes region unit open or shut. AN outsized separation relates to Eye conclusion. On the off chance that the eyes region unit discovered shut for 5 back to back edges, the framework draws in the Conclusion that the main impetus is nodding off and gives an alarm. The framework is likewise ready to watch once the eyes can not be found, and works beneath modest lighting Conditions. here we will in general also track client live area on the off chance that any crisis shows up, at that point framework precisely send area to closest emergency clinic, police central command comparatively its individuals from the family. right now we will in general also notice client square measure alcoholic or not by abuse liquor police work sensors. Here we will say that our framework is extra affordable that current frameworks. Shruti Chandrakant Zarekar | Priyanka Dattatray Desai | Manjusha Prabhakar Randhave| Surajsingh Rajendrasingh Chauhan | Dr. Shyam Gupta "Drowsiness and Alcohol Detection for Accident Prevention using Machine Learning" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-4 | Issue-6 , October 2020, URL: https://www.ijtsrd.com/papers/ijtsrd33587.pdf Paper Url: https://www.ijtsrd.com/engineering/computer-engineering/33587/drowsiness-and-alcohol-detection-for-accident-prevention-using-machine-learning/shruti-chandrakant-zarekar
IRJET - Prediction of Autistic Spectrum Disorder based on Behavioural Fea...IRJET Journal
This document summarizes a research paper that aims to predict autism spectrum disorder (ASD) based on behavioral features using machine learning. The researchers collected ASD screening data from different age groups to develop and evaluate neural network models for predicting ASD. They achieved up to 90% accuracy in predicting ASD. The researchers concluded that machine learning is a promising approach for ASD prediction but noted limitations like lack of large datasets. They plan to improve the models by collecting more data from various sources.
This document summarizes a research paper that proposes a non-intrusive system to detect driver drowsiness in real-time using eye closure ratio as an input. The system uses a Raspberry Pi camera to capture photos of the driver's eyes. If the eye closure ratio decreases below a standard ratio, the driver is alerted via notification and an email is sent to the vehicle owner. The paper describes the proposed system methodology in detail using data flow diagrams, flowcharts, and descriptions of the facial landmark detection and eye aspect ratio algorithms used to detect drowsiness.
The document describes a proposed driver drowsiness detection system that uses computer vision techniques. The system aims to accurately monitor the open and closed state of a driver's eyes by analyzing a sequence of images of the eyes and face. This will allow the system to detect early symptoms of driver fatigue in order to avoid accidents. Existing approaches are discussed along with their disadvantages. The proposed system uses a camera to detect the face and eyes, and will detect fatigue and provide alarms or notifications by analyzing eye movements and markings over time. It will have advantages over sensor-based techniques by providing driver assistance without needing direct physical contact.
This is a summary of Siegel's Predictive Analytics. The presentation encourages the reader to actively look at the capabilities of Predictive Analytics as a tool to make a forecast for one entity.
This document summarizes the results of a survey of 81 refractive surgeons on the business impacts of adopting the Intralase femtosecond laser for LASIK flap creation. Key findings include:
1) Practices typically increased LASIK fees by an average of $394 after adopting Intralase to offset costs. Higher fee practices saw smaller increases.
2) Over 80% of patients chose Intralase when given the option, leading more surgeons to make it standard.
3) Practices saw an average 41% increase in LASIK revenue and a 40% decrease in retreatments after adopting Intralase.
This document summarizes medical imaging technologies developed by the Institute for Infocomm Research for screening major ocular diseases in primary care. It describes technologies for automatically grading cataracts, calculating cup-to-disc ratios for glaucoma screening, and detecting various eye conditions like drusen, pathological myopia, and cornea diseases from different image modalities like lens, fundus, OCT, and RETCAM images. It also details the AGLAIA system for automatic glaucoma diagnosis and its validation in a study of 1,676 subjects where it achieved diagnostic performance comparable to expert grading of fundus photos and eye pressure measurements.
IRJET- GDPS - General Disease Prediction SystemIRJET Journal
The document describes a General Disease Prediction System (GDPS) that uses machine learning and data mining techniques to predict diseases based on patient symptoms.
The GDPS first collects patient data, preprocesses it, and extracts relevant features. It then implements the ID3 decision tree algorithm to generate a predictive model and classify diseases. As an admin, one can train the model using sample data. As a user, one can enter symptoms and the trained model will predict the likely disease and recommend precautions.
The GDPS was tested on a dataset of 120 patients and achieved 86.67% accuracy in disease prediction. The system currently covers common diseases but future work involves expanding it to predict more serious or fatal diseases like various cancers
Credit card fraud detection through machine learningdataalcott
This document discusses using machine learning algorithms for credit card fraud detection. It proposes using principal component analysis for feature selection followed by logistic regression and decision tree models. It finds that logistic regression has higher accuracy at 79.91% compared to 71.41% for decision tree. The proposed approach aims to better handle imbalanced data and reduce fraudulent transactions. Future work could implement the approach in Python and produce experimental results.
LAN Based HF Radio Simulator An Approach to Develop an Early Prototype for Lo...YogeshIJTSRD
This document summarizes the key aspects of developing a LAN-based HF radio simulator. It conducted surveys and interviews with users to understand requirements. The surveys found that a simulator could facilitate training by allowing practice without actual radios. A web-based platform was preferred. The document outlines the methodology, which includes specifying user and software requirements. It defines functional and non-functional requirements for the simulator. The simulator aims to allow trainees to practice operating radios in different scenarios while preserving performance records.
A New Proposal for Smartphone Based Drowsiness Detection and Warning System f...YogeshIJTSRD
This paper proposes a Smartphone based system for the detection of drowsiness in automotive drivers. The proposed system uses three stage drowsiness detection technique. The first stage uses the percentage of eyelid closure PERCLOS which is obtained by capturing images with the front camera of the Smartphone with a modified eye state classification method. The system uses near infrared lighting for illuminating the face of the driver during night driving. The second step uses the voiced to the unvoiced ratio VUR obtained from the speech data from the microphone, in the event PERCLOS crosses the threshold. The VUR is also compared with a threshold and if it is a value greater than that of the threshold, it moves on to the next verification stage. In the final verification stage, touch response is required within the stipulated time to declare whether the driver is drowsy or not and subsequently sound an alarm. To awake the driver, a vibrating mechanism is done and also the live GPS location is also sent to an emergency contact. We have studied eight other reference papers for the literature review. The system has three advantages over existing drowsiness detection systems. First, the three stage verification process makes the system more reliable. The second advantage is its implementation on an Android smart phone, which is readily available to most drivers or cab owners as compared to other general purpose embedded platforms. The third advantage is the use of SMS service to inform the control room as well as the passenger regarding the loss of attention of the driver. Abishek K Biju | Godwin Jolly | Asif Mohammed C A | Dr. Paul P Mathai | Derek Joseph "A New Proposal for Smartphone-Based Drowsiness Detection and Warning System for Automotive Drivers" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-5 | Issue-5 , August 2021, URL: https://www.ijtsrd.com/papers/ijtsrd45083.pdf Paper URL: https://www.ijtsrd.com/engineering/computer-engineering/45083/a-new-proposal-for-smartphonebased-drowsiness-detection-and-warning-system-for-automotive-drivers/abishek-k-biju
This document summarizes research on driver drowsiness detection systems. It begins with an abstract describing how convolutional neural networks were used to create a model to classify eyes as open or closed using webcam photos. It then reviews six related works that used techniques like LSTM-CNN architectures, support vector machines, and multi-modal information to detect drowsiness. The document outlines the objectives, methodology, application requirements, and conclusions of the research, which involved using image processing and deep learning models to accurately detect drowsiness in a non-invasive manner to improve road safety.
An efficient system for real time fatigue detectionAlexander Decker
This document summarizes a research paper that proposes an efficient system for real-time fatigue detection. The system uses computer vision and image processing techniques to measure eye closure count, blinking rate, and yawning to detect user fatigue. Face detection is performed using the Viola-Jones algorithm. Abnormalities in eye and mouth behavior are then analyzed to determine if the user is fatigued. The system aims to detect fatigue early enough to avoid accidents in applications where user attentiveness is critical. It is designed to have low time and space complexity, be low cost, and not significantly impact normal user interactions. The proposed approach and algorithm are described, and example results of fatigue detection are provided.
A presentation of Driver drowsiness alert system which can identify whether the driver is attentive or sleepy while driving and hence alert them by a beep when the driver is sleepy.Python and open CV are main technologies used here along with hass cascade algorithm for the same.
Driving without license is the major cause for the road accident and the equivalent monetary losses. This paper is based on virtual reality based driving system which would enhance road safety and vehicle security. This paper helps to limit the vehicle operation on the basics of two parameters-Learn the driving by our own, category (car or bike) of the vehicle for which the driving license is issued. The hardware and software system required to improve our safety and security is developed. This driving system is apt for getting the license without bribe by gathering eye-gaze, Electroencephalography and peripheral physiological data.
When driving long distances, drivers who do not take frequent rests are more likely to get sleepy, a condition that experts say they often fail to identify early enough. Based on eye condition, this research proposes a system for detecting driver sleepiness in real time. A camera is often used to take a sequence of images by the system. In our system, these capture images may be saved as individual frames. The resulting frame is sent into facial recognition software as an input. The image's needed feature (eye) is then extracted. The method creates a condition for each eye and suggests a certain number of frames with the same condition that can be registered.
Driver's drowsiness is the main reason for vehicular accidents. Drowsy driving is the form of impaired driving
that continuously affects a person's ability to drive safely. Continuous restless driving for longer time may result in
drowsiness and cause accidents. In this study, a collaborative system is build which assist the user and identifies
his/her state while driving in order to improve safety by preventing accidents. Based on grayscale image processing,
the position of the driver's face and his/her head movement is analysed. The driver's state identification also includes
the detection of alcohol consumption with the help of sensors.
Enhancing Time Series Anomaly Detection: A Hybrid Model Fusion ApproachIJCI JOURNAL
Exploring and Identifying Anomalies in time-series data is very crucial in today’s world revolve around data. These data are being used to make important decisions; hence, an efficient and reliable anomaly detection system should be involved in this process to ensure that the best decisions are being made. The paper explores other types of anomalies and proposes efficient detection methods which can be used. Anomalies are patterns that deviate from usual expected behavior. These can come from system failures or unexpected activity. This research paper explores the vulnerabilities of commonly used anomaly detection algorithms such as the Z-Score and static threshold approach. Each method used in this paper has its unique capabilities and limitations. These methods range from using statistical methods and machine learning approaches to detecting anomalies in a time-series dataset. Furthermore, this paper explores other open-source libraries that can be used to detect anomalies, such as Greykite and Prophet Python library. This paper serves as a good source for anyone new to anomaly detection and willing to explore.
Driver Drowsiness is a grave issue resulting in many road accidents each year. To evaluate the exact number of sleep related accidents because of the difficulties in detecting whether fatigue was a factor and in assessing the level of fatigue is not currently possible. In this paper the camera will be placed besides the rare view mirror of car in way such that it is in clear view of the frontal face of the driver. This camera will continuously capture the video of driver’s frontal face while driving. The system will detect the frontal face in the image and later the eyes. Depending upon the conditions the system will generate an alert. The focus will be on the system that will accurately monitor the open or closed state of the driver’s eyes in real-time. By monitoring the eyes, it is believed that the symptoms of driver fatigue can be detected early to avoid accidents.
There are different types of algorithms used in eye tracking technologies. These algorithms are divided into two main categories: feature-based and model-based. Feature-based technologies consist of threshold values, which are used to decide the presence or absence of features or determinant factors. While the model-based approach is an iterative search of a model parameter, which is the best fitting model that is a closest match to the image. However, these approaches have significant problems regarding computational speed and accuracy. Similarly, there are different types of eye – tracking technologies, which depend on different types of technologies such as infrared video cameras and other technologies, which require specific calibration and setup and are quite expensive. Therefore, in this paper, we propose an alternative eye–tracking technology using a new eye-tracking algorithm, which is highly portable and independent of any hardware or software systems. In an evaluation the algorithm worked accurately for users with strong dyslexia. Participants had various positive and negative opinions regarding such an auto-detection system. Furthermore, we propose that such technology could be used to automatically modify the content of online material to better suit dyslexic users.
AN ALGORITHM FOR AUTOMATICALLY DETECTING DYSLEXIA ON THE FLYijcsit
There are different types of algorithms used in eye tracking technologies. These algorithms are divided into two main categories: feature-based and model-based. Feature-based technologies consist of threshold values, which are used to decide the presence or absence of features or determinant factors. While the model-based approach is an iterative search of a model parameter, which is the best fitting model that is a
closest match to the image. However, these approaches have significant problems regarding computational speed and accuracy. Similarly, there are different types of eye – tracking technologies, which depend on different types of technologies such as infrared video cameras and other technologies, which require specific calibration and setup and are quite expensive. Therefore, in this paper, we propose an alternative eye–tracking technology using a new eye-tracking algorithm, which is highly portable and independent of any hardware or software systems. In an evaluation the algorithm worked accurately for users with strong dyslexia. Participants had various positive and negative opinions regarding such an auto-detection system. Furthermore, we propose that such technology could be used to automatically modify the content of online material to better suit dyslexic users.
Drowsiness and Alcohol Detection for Accident Prevention using Machine Learningijtsrd
A Drowsy Driver Detection System has been created, utilizing a non meddling machine Vision based absolutely ideas. The framework utilizes a touch monochrome surveillance camera that focuses Directly towards the drivers face and screens the drivers eyes along these lines on watch weakness. In Such a case once exhaustion is identified, an alarm is given to caution the main impetus. This Report depicts the gratitude to see the eyes, and together the gratitude to check if the eyes zone unit open or Closed. The algorithmic standard created is restrictive to any directly unconcealed papers, that was a Primary goal of the venture. The framework manages exploitation data acquired for the Binary form of the picture to go glancing out the edges of the face, that limits the domain of where the Eyes may exist. When the face region is discovered, the eyes zone unit found by registering the flat Averages at stretches the region. Taking into thought the data that eye locales at stretches the face blessing decent power changes, the eyes zone unit put by finding the various force changes at spans the face. When the eyes zone unit set, live the separations between the force changes at spans the consideration zone confirm whether or not the eyes region unit open or shut. AN outsized separation relates to Eye conclusion. On the off chance that the eyes region unit discovered shut for 5 back to back edges, the framework draws in the Conclusion that the main impetus is nodding off and gives an alarm. The framework is likewise ready to watch once the eyes can not be found, and works beneath modest lighting Conditions. here we will in general also track client live area on the off chance that any crisis shows up, at that point framework precisely send area to closest emergency clinic, police central command comparatively its individuals from the family. right now we will in general also notice client square measure alcoholic or not by abuse liquor police work sensors. Here we will say that our framework is extra affordable that current frameworks. Shruti Chandrakant Zarekar | Priyanka Dattatray Desai | Manjusha Prabhakar Randhave| Surajsingh Rajendrasingh Chauhan | Dr. Shyam Gupta "Drowsiness and Alcohol Detection for Accident Prevention using Machine Learning" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-4 | Issue-6 , October 2020, URL: https://www.ijtsrd.com/papers/ijtsrd33587.pdf Paper Url: https://www.ijtsrd.com/engineering/computer-engineering/33587/drowsiness-and-alcohol-detection-for-accident-prevention-using-machine-learning/shruti-chandrakant-zarekar
IRJET - Prediction of Autistic Spectrum Disorder based on Behavioural Fea...IRJET Journal
This document summarizes a research paper that aims to predict autism spectrum disorder (ASD) based on behavioral features using machine learning. The researchers collected ASD screening data from different age groups to develop and evaluate neural network models for predicting ASD. They achieved up to 90% accuracy in predicting ASD. The researchers concluded that machine learning is a promising approach for ASD prediction but noted limitations like lack of large datasets. They plan to improve the models by collecting more data from various sources.
This document summarizes a research paper that proposes a non-intrusive system to detect driver drowsiness in real-time using eye closure ratio as an input. The system uses a Raspberry Pi camera to capture photos of the driver's eyes. If the eye closure ratio decreases below a standard ratio, the driver is alerted via notification and an email is sent to the vehicle owner. The paper describes the proposed system methodology in detail using data flow diagrams, flowcharts, and descriptions of the facial landmark detection and eye aspect ratio algorithms used to detect drowsiness.
The document describes a proposed driver drowsiness detection system that uses computer vision techniques. The system aims to accurately monitor the open and closed state of a driver's eyes by analyzing a sequence of images of the eyes and face. This will allow the system to detect early symptoms of driver fatigue in order to avoid accidents. Existing approaches are discussed along with their disadvantages. The proposed system uses a camera to detect the face and eyes, and will detect fatigue and provide alarms or notifications by analyzing eye movements and markings over time. It will have advantages over sensor-based techniques by providing driver assistance without needing direct physical contact.
IRJET - Real Time Facial Analysis using Tensorflowand OpenCVIRJET Journal
This document presents a real-time facial analysis system using TensorFlow and OpenCV. The system can detect facial expressions, age, and gender from images and video in real-time. It uses deep learning models trained on facial datasets to analyze faces. The system is designed for applications like security, attendance tracking, and finding lost children. It works by extracting facial features from images, applying preprocessing techniques, classifying faces, and making predictions about attributes. The document discusses the methodology, existing techniques like PCA and HMM, the proposed system architecture, sample code, and conclusions.
Driver Drowsiness Detection System Using Image ProcessingIRJET Journal
This document describes a driver drowsiness detection system using image processing. The system uses a Raspberry Pi camera to capture video frames of the driver's face. It then uses computer vision techniques like Haar cascade classifiers and the Hough transform to detect and track the face and eyes in each frame. It monitors the number of consecutive frames where the eyes are closed to determine if the driver is drowsy. If the eyes are closed for more than 2 seconds, it will trigger an alarm. The system aims to reduce road accidents caused by driver fatigue by providing an early warning of drowsiness without any physical probes contacting the driver.
DRIVER DROWSINESS DETECTION USING DEEP LEARNINGIRJET Journal
This document presents a driver drowsiness detection system using deep learning. It begins with an introduction describing the safety issues caused by drowsy driving and the need for such a system. It then discusses the proposed system which uses a CNN trained on eye image data to classify eyes as open or closed in real-time video. If the eyes are classified as closed for a certain number of frames, an alert is triggered. The system achieved 96% accuracy on a test dataset. It concludes that CNNs provide better performance than other facial extraction methods for drowsiness detection.
IRJET- Analysis of Yawning Behavior in IoT based of Drowsy DriversIRJET Journal
This document presents a system for detecting driver drowsiness based on analyzing yawning behavior using IoT sensors and devices. The system would detect yawns using both geometric features of the mouth and eyes as well as appearance-based features. Yawns would be detected even if the mouth is covered by a hand. The goal is to incorporate yawn detection into a hybrid drowsiness detection system to help reduce accidents caused by fatigued driving. The proposed approach was found to detect both uncovered and covered yawns with 95% accuracy.
Similar to Detecting Fatigue Driving Through PERCLOS: A Review (20)
QA or the Highway - Component Testing: Bridging the gap between frontend appl...zjhamm304
These are the slides for the presentation, "Component Testing: Bridging the gap between frontend applications" that was presented at QA or the Highway 2024 in Columbus, OH by Zachary Hamm.
Northern Engraving | Nameplate Manufacturing Process - 2024Northern Engraving
Manufacturing custom quality metal nameplates and badges involves several standard operations. Processes include sheet prep, lithography, screening, coating, punch press and inspection. All decoration is completed in the flat sheet with adhesive and tooling operations following. The possibilities for creating unique durable nameplates are endless. How will you create your brand identity? We can help!
In the realm of cybersecurity, offensive security practices act as a critical shield. By simulating real-world attacks in a controlled environment, these techniques expose vulnerabilities before malicious actors can exploit them. This proactive approach allows manufacturers to identify and fix weaknesses, significantly enhancing system security.
This presentation delves into the development of a system designed to mimic Galileo's Open Service signal using software-defined radio (SDR) technology. We'll begin with a foundational overview of both Global Navigation Satellite Systems (GNSS) and the intricacies of digital signal processing.
The presentation culminates in a live demonstration. We'll showcase the manipulation of Galileo's Open Service pilot signal, simulating an attack on various software and hardware systems. This practical demonstration serves to highlight the potential consequences of unaddressed vulnerabilities, emphasizing the importance of offensive security practices in safeguarding critical infrastructure.
Northern Engraving | Modern Metal Trim, Nameplates and Appliance PanelsNorthern Engraving
What began over 115 years ago as a supplier of precision gauges to the automotive industry has evolved into being an industry leader in the manufacture of product branding, automotive cockpit trim and decorative appliance trim. Value-added services include in-house Design, Engineering, Program Management, Test Lab and Tool Shops.
Lee Barnes - Path to Becoming an Effective Test Automation Engineer.pdfleebarnesutopia
So… you want to become a Test Automation Engineer (or hire and develop one)? While there’s quite a bit of information available about important technical and tool skills to master, there’s not enough discussion around the path to becoming an effective Test Automation Engineer that knows how to add VALUE. In my experience this had led to a proliferation of engineers who are proficient with tools and building frameworks but have skill and knowledge gaps, especially in software testing, that reduce the value they deliver with test automation.
In this talk, Lee will share his lessons learned from over 30 years of working with, and mentoring, hundreds of Test Automation Engineers. Whether you’re looking to get started in test automation or just want to improve your trade, this talk will give you a solid foundation and roadmap for ensuring your test automation efforts continuously add value. This talk is equally valuable for both aspiring Test Automation Engineers and those managing them! All attendees will take away a set of key foundational knowledge and a high-level learning path for leveling up test automation skills and ensuring they add value to their organizations.
Essentials of Automations: Exploring Attributes & Automation ParametersSafe Software
Building automations in FME Flow can save time, money, and help businesses scale by eliminating data silos and providing data to stakeholders in real-time. One essential component to orchestrating complex automations is the use of attributes & automation parameters (both formerly known as “keys”). In fact, it’s unlikely you’ll ever build an Automation without using these components, but what exactly are they?
Attributes & automation parameters enable the automation author to pass data values from one automation component to the next. During this webinar, our FME Flow Specialists will cover leveraging the three types of these output attributes & parameters in FME Flow: Event, Custom, and Automation. As a bonus, they’ll also be making use of the Split-Merge Block functionality.
You’ll leave this webinar with a better understanding of how to maximize the potential of automations by making use of attributes & automation parameters, with the ultimate goal of setting your enterprise integration workflows up on autopilot.
The Microsoft 365 Migration Tutorial For Beginner.pptxoperationspcvita
This presentation will help you understand the power of Microsoft 365. However, we have mentioned every productivity app included in Office 365. Additionally, we have suggested the migration situation related to Office 365 and how we can help you.
You can also read: https://www.systoolsgroup.com/updates/office-365-tenant-to-tenant-migration-step-by-step-complete-guide/
AI in the Workplace Reskilling, Upskilling, and Future Work.pptxSunil Jagani
Discover how AI is transforming the workplace and learn strategies for reskilling and upskilling employees to stay ahead. This comprehensive guide covers the impact of AI on jobs, essential skills for the future, and successful case studies from industry leaders. Embrace AI-driven changes, foster continuous learning, and build a future-ready workforce.
Read More - https://bit.ly/3VKly70
Getting the Most Out of ScyllaDB Monitoring: ShareChat's TipsScyllaDB
ScyllaDB monitoring provides a lot of useful information. But sometimes it’s not easy to find the root of the problem if something is wrong or even estimate the remaining capacity by the load on the cluster. This talk shares our team's practical tips on: 1) How to find the root of the problem by metrics if ScyllaDB is slow 2) How to interpret the load and plan capacity for the future 3) Compaction strategies and how to choose the right one 4) Important metrics which aren’t available in the default monitoring setup.
"Frontline Battles with DDoS: Best practices and Lessons Learned", Igor IvaniukFwdays
At this talk we will discuss DDoS protection tools and best practices, discuss network architectures and what AWS has to offer. Also, we will look into one of the largest DDoS attacks on Ukrainian infrastructure that happened in February 2022. We'll see, what techniques helped to keep the web resources available for Ukrainians and how AWS improved DDoS protection for all customers based on Ukraine experience
[OReilly Superstream] Occupy the Space: A grassroots guide to engineering (an...Jason Yip
The typical problem in product engineering is not bad strategy, so much as “no strategy”. This leads to confusion, lack of motivation, and incoherent action. The next time you look for a strategy and find an empty space, instead of waiting for it to be filled, I will show you how to fill it in yourself. If you’re wrong, it forces a correction. If you’re right, it helps create focus. I’ll share how I’ve approached this in the past, both what works and lessons for what didn’t work so well.
Conversational agents, or chatbots, are increasingly used to access all sorts of services using natural language. While open-domain chatbots - like ChatGPT - can converse on any topic, task-oriented chatbots - the focus of this paper - are designed for specific tasks, like booking a flight, obtaining customer support, or setting an appointment. Like any other software, task-oriented chatbots need to be properly tested, usually by defining and executing test scenarios (i.e., sequences of user-chatbot interactions). However, there is currently a lack of methods to quantify the completeness and strength of such test scenarios, which can lead to low-quality tests, and hence to buggy chatbots.
To fill this gap, we propose adapting mutation testing (MuT) for task-oriented chatbots. To this end, we introduce a set of mutation operators that emulate faults in chatbot designs, an architecture that enables MuT on chatbots built using heterogeneous technologies, and a practical realisation as an Eclipse plugin. Moreover, we evaluate the applicability, effectiveness and efficiency of our approach on open-source chatbots, with promising results.
Introducing BoxLang : A new JVM language for productivity and modularity!Ortus Solutions, Corp
Just like life, our code must adapt to the ever changing world we live in. From one day coding for the web, to the next for our tablets or APIs or for running serverless applications. Multi-runtime development is the future of coding, the future is to be dynamic. Let us introduce you to BoxLang.
Dynamic. Modular. Productive.
BoxLang redefines development with its dynamic nature, empowering developers to craft expressive and functional code effortlessly. Its modular architecture prioritizes flexibility, allowing for seamless integration into existing ecosystems.
Interoperability at its Core
With 100% interoperability with Java, BoxLang seamlessly bridges the gap between traditional and modern development paradigms, unlocking new possibilities for innovation and collaboration.
Multi-Runtime
From the tiny 2m operating system binary to running on our pure Java web server, CommandBox, Jakarta EE, AWS Lambda, Microsoft Functions, Web Assembly, Android and more. BoxLang has been designed to enhance and adapt according to it's runnable runtime.
The Fusion of Modernity and Tradition
Experience the fusion of modern features inspired by CFML, Node, Ruby, Kotlin, Java, and Clojure, combined with the familiarity of Java bytecode compilation, making BoxLang a language of choice for forward-thinking developers.
Empowering Transition with Transpiler Support
Transitioning from CFML to BoxLang is seamless with our JIT transpiler, facilitating smooth migration and preserving existing code investments.
Unlocking Creativity with IDE Tools
Unleash your creativity with powerful IDE tools tailored for BoxLang, providing an intuitive development experience and streamlining your workflow. Join us as we embark on a journey to redefine JVM development. Welcome to the era of BoxLang.
Introducing BoxLang : A new JVM language for productivity and modularity!
Detecting Fatigue Driving Through PERCLOS: A Review
1. Samuel Kim, Irfan Wisanggeni, Ryan Ros & Rania Hussein
International Journal of Image Processing (IJIP), Volume (14) : Issue (1) : 2020 1
Detecting Fatigue Driving Through PERCLOS: A Review
Samuel Kim syjkim75@uw.edu
Department of Electrical Engineering
University of Washington
Seattle, 98195, United States of America
Irfan Wisanggeni irfanw@uw.edu
Department of Electrical Engineering
University of Washington
Seattle, 98195, United States of America
Ryan Ros rosr99@uw.edu
Department of Electrical Engineering
University of Washington
Seattle, 98195, United States of America
Rania Hussein rhussein@uw.edu
Department of Electrical Engineering
University of Washington
Seattle, 98195, United States of America
Abstract
In this paper, we present a literature survey about drowsy driving detection using PERCLOS
metric that determines the percentage of eye closure. This metric determines that an eye is
closed if the percentage of eye closure is 80% or above. When this percentage is observed for
multiple frames of a video camera feed, the driver is determined to be in an unsafe fatigue status.
In our research, we found that the PERCLOS metric had a 0.79 to 0.87 correlation coefficient
value which exceeds the 0.7 R value needed to be considered a strong correlation coefficient. A
higher value than 0.7 indicates a more linear relationship which means that the metric is
dependable [1].
Keywords: PERCLOS, Real-time Systems, Autonomous Driving.
1. INTRODUCTION
In this modern age, many people go to work, school, and other places using their transportation.
According to a 2018 census from Statista Research Department, there is a recorded 276 million
cars in the United States. While public transportation use may be increasing, many Americans
certainly use cars as their main form of transportation. Since there are many cars on the road, the
chances of being involved in a car accident are high. Although many motor accidents occur year-
round, there is one cause that reoccurs, drowsy driving, and the rate of these accidents is steadily
increasing. According to the National Highway Traffic Safety Administration (NHTSA), there were
an estimated 91,000 accidents, 50,000 people injured, and 800 deaths due to drowsy driving in
2017 alone. However, it is agreed by sleep science, traffic safety, and other health communities
that the number of accidents is an underestimate of the people injured or killed by drowsy driving.
Some say that there could be potentially 6000 fatal car crashes. From this information, it is clear
there is a need for ways to prevent drowsy driving. These dangers have stemmed from research
on creating the most effective way to alert the driver when they are detected to be drowsy. After
looking at numerous scientific papers on the topic of drowsy driving, the topic for this survey
paper was narrowed down to drowsy driving detection through the use of eye detection and blink
counting, particularly the ones that employ the use of PERCLOS, which is an eye detection
2. Samuel Kim, Irfan Wisanggeni, Ryan Ros & Rania Hussein
International Journal of Image Processing (IJIP), Volume (14) : Issue (1) : 2020 2
method that involves detecting when the eye is at least a certain percentage closed, most cases
being 80% closed from its normal state.
The reason behind this survey paper is to answer the question of whether drowsy driving
detection with eye and blink detection was viable in widespread use. To gauge the effectiveness
of the system, we were interested in both the success rate of the system, and whether it can be
implemented with as little hardware as possible. These two factors were highly considered when
we were searching for research papers. After we parsed through the papers, we were able to find
the similarities and differences between each implementation of a drowsy detection system. The
main thing we wanted to find out was if it was possible to create a system with minimal hardware
that could have a success rate of 80% or higher. Throughout this paper, we will discuss the
similarities, differences and the overall usefulness of each implementation.
2. PROBLEM STATEMENT
Is it possible to create a minimal hardware implementation with a success rate of 80% or over?
This was the question statement we wanted to answer when writing this paper. Just from initial
screening of the research papers we looked through, we came up with a hypothesis that it is
possible to create a system that has minimal hardware and functions with an 80% or over
succession rate. We came up with this hypothesis because we were able to find evidence that
other implementations have successfully created the system that we seek. To prove that a
system exists, we will go over the evidence we have found.
3. CHARACTERIZATON OF CLASSES
We brought up that the two key factors to determine if the system is valid or not were the success
rate of the system, and the cost of the implementation.
3.1 Class 1
One of the more popular papers that fit these criteria was titled “Real-time nonintrusive monitoring
and prediction of driver fatigue” by researchers Q. Ji, Z. Zhu, and P. Lan [2]. Though this is not
the most recent paper out there, it has been cited in 358 other papers as of the writing of this
survey paper. Their system focused on creating a way to predict driver fatigue in a way that was
non-intrusive and involved using two cameras, outfitted with infrared LEDs to brighten the driver’s
face and make eye detection/eye tracking easier. Through a validation process by the research
team, they were able to determine the success rate of approximately 95.75% with 0.05% of the
failure coming from false-positives and 4.2% coming from missed readings.
3.2 Class 2
The next research paper is an even older one that also has several citations titled “A drowsy
driver detection system for heavy vehicles” by R. Grace, V. Byrne, D. Bierman, J. Legrand, D.
Gricourt, R. Davis, J. Staszewski, and B. Carnahan [3]. This paper was written way back in
November of 1998 and has been cited in 49 research papers and 24 patents. As this was an
older paper, the technology available to them was limited. The camera that this research team
had used had a maximum frame rate of 6 frames per second and was said to correlate “very
highly” with manually coded data and had a high repeatability given the same datasets.
3.3 Class 3
The next research paper that we had found to fit our criteria was a paper titled “A method of
driving fatigue detection based on eye location” written by L. Li, M. Xie, and H. Dong [4]. This
paper introduced a method to locate the eyes based on Active Appearance Model (AAM). This
system consists of narrowing on the eyes through different filters and AAM. It first uses an
Adaboost algorithm module to get a rough location of the face and eyes. It them uses the AAM
model to match the image to the training samples and accurately locate certain features of the
eyes. Once the features of the eyes are known, the PERCLOS score can be calculated. This
paper goes on to explain some of the limitations that may occur with a system that involves only a
camera since things such as lighting, movement of the driver’s head, or eyewear can throw of the
3. Samuel Kim, Irfan Wisanggeni, Ryan Ros & Rania Hussein
International Journal of Image Processing (IJIP), Volume (14) : Issue (1) : 2020 3
detection. The researchers of this paper did not include the success rate of their system, but the
hardware used, being only the camera to record the driver, did not seem expensive.
3.4 Class 4
M. Poursadeghiyan [5] had a research paper that fit our criteria. According to their research
paper, “Using Image Processing in the Proposed Drowsiness Detection System”, they were able
to create a system that had a success rate of 93% which is high above the requirement that we
set as a baseline. The implementation used a camera to take a video of the subject. Each frame
in the video would be checked for the coordination of facial details. Using the Viola-Jones
algorithm, each frame of the video would be checked for specific facial landmarks of the eyes of
the driver. To find out if the driver’s eyes were closed or not, the frame would be converted to
grayscale and checked to see if eyes are closed or not. We then checked to see how they
implemented their system. According to Poursadeghiyan [5], the system they used required
extensive equipment including a virtual driver simulator, computer, and camera. Indeed, the
system worked with high success, but it couldn’t meet our mark of using minimal hardware. Aside
from the simulator used to conduct the research, the computer and camera system they used
requires more space than other systems. Because of this, we continued to look for other
implementations that could satisfy our two requirements.
3.5 Class 5
The next paper we looked at was by C. Xu [6] and it was cited by 5 other papers along with a
patent. According to “Efficient eye detection in real-time for drowsy driving monitoring system”, Xu
[6] was able to create a system by using a Local Binary Pattern histogram to create a histogram
of the eye region and then passed through an Adaboost cascade classifier. This classifier would
be trained by the numerous histograms and it would be able to determine if the eye was closed or
not according to a PERCLOS score. After looking at this paper, we were able to identify that the
success rate of this implementation was averaging 98% success. This system’s success rate was
way above the requirement we set so we wanted to see what kind of equipment was being used
to create this system. We noticed that for this setup, it was not explicitly stated that it could have
been created with minimal hardware.
3.6 Class 6
According to “Sober-Drive: A smartphone-assisted drowsy driving detection system”, L. Xu [7]
used a system that leveraged the PERCLOS metric to train a neural network that could identify if
the driver was drowsy or not. The more important factor was that for the first time looking through
our papers, we were able to find a system that could be implemented with minimal hardware.
Xu’s [7] implementation could run on an android phone without any other equipment. Also,
according to this research paper, the success rate of their implementation was averaging a 90%
success rate. This implementation system satisfied both factors we set at the beginning of our
paper. The success rate was above 80% and it could function with minimal hardware, which for
this system was using an android phone. Xu’s team's implementation used a neural network to
determine the drowsiness of the driver.
3.7 Class 7
According to “Efficient Measurement of Eye Blinking under Various Illumination Conditions for
Drowsiness Detection Systems”, I. Park [8] was able to implement a system using a standard 2D
camera plus two IR illuminators and a computing power of some sort. As stated by the paper,
Park’s team was able to have an average success rate of 94% without illumination compensation
and an average success rate of 98% with illumination compensation. Both success rates far
exceeded the baseline of an 80% success rate. The second factor was somewhat satisfied as the
system itself doesn’t require as much hardware as other implementations, but it is not the
smallest it could be. The software that was created was like the first two systems that are
previously discussed above.
4. Samuel Kim, Irfan Wisanggeni, Ryan Ros & Rania Hussein
International Journal of Image Processing (IJIP), Volume (14) : Issue (1) : 2020 4
3.8 Class 8
The last system we looked at was the best system that we found since it exceeded our
expectations for both the success rate and the use of minimum hardware. Professor Y. Cheung
and his fellow researchers created a system that would detect the eye region of the driver’s face
and find the distance between the eye corner and the iris to detect whether the driver was falling
asleep or not. According to “US Patent: US9,563,805” [9], Cheung’s system was able to have an
average success rate of 93%. Cheung’s system was also implemented in a cell phone, so it
passed the criteria of using minimum hardware. This system was the most effective system out of
all the ones that we found in our sources. The combination of having a high success rate plus
being able to run it on low-end cell phone hardware stood out as being overall the most effective
implementation for detecting drowsy driving.
Techniques for Drowsy Driving Detection Classes
Using multiple cameras includes use of multiple
2d camera, IR camera, depth sensing camera.
1, 7
Using single camera 2, 3, 4, 5, 6, 8
Using neural network to compute different
scenario
3, 4, 5, 6
Detection through distance measurement 8
Mobile hardware implementation 6, 8
Desktop hardware implementation 1, 2, 3, 4, 5, 7
4. COMPARATIVE EVALUATION
In the previous section, we were able to categorize each class with certain categories. It was
common to see overlap between each of the classes but at the same time, we were able to
identify clear differences. In this section, we want to discuss how each system compares to each
other and possible changes to make a hybrid system. We will back up our recommendation
based on the chart provided above. Looking at the first category, it was surprising to see that not
many implementations used multiple cameras to detect drowsy driving. However, there was two
implementations that used multiple cameras. However, the success rate was not much higher
than other implementations. The next technique was using only one camera for the overall
system. This was much more common, and every single implementation had a success rate of
over 80%, so having one camera wasn’t a disadvantage. It is an advantage because those
systems would require less hardware to make their system work. Our recommendation for
between these two techniques is to use a system with one camera since the success rate is
around the same as other multiple camera implementations and it would require less equipment.
The next two techniques dealt with how each system identified a drowsy driver. The first
technique was to use a machine learning network that was trained with images to differentiate
between awake and drowsy drivers. In practice, the implementations that used neural
networks/machine learning tended to have higher success rates than the ones that used distance
measurements from camera data. The second technique was to directly use the camera data to
find if the driver was drowsy or not. For example, packages such as OpenCV, have methods that
can give researchers data to calculate the distance that is used to identify if the driver is drowsy
or not. This method will exceed the minimum 80% success rate and require less memory
because it doesn’t need image data to compare too. For our recommendation, we would choose
to use the machine learning with one exception. Researchers should process and store the data
necessary for the neural network on the cloud so that local memory allocated to the device is to a
minimum. This is important for the next technique. The last set of techniques deals with how the
hardware is packaged. The first technique is to use a mobile implementation, more specifically, a
mobile phone. The benefits to using a mobile phone as a computing device is that the access to
smartphones in this decade has been more readily available than it has in the past. With more
mobile phones, providing an application can be better for more widespread use compared to
proprietary hardware. The second technique is to use desktop hardware which could provide
more computing power at the cost of portability. Between these two techniques, we recommend
using the mobile implementation as it provides the most ideal power to portability ratio. To recap,
5. Samuel Kim, Irfan Wisanggeni, Ryan Ros & Rania Hussein
International Journal of Image Processing (IJIP), Volume (14) : Issue (1) : 2020 5
we would recommend creating a mobile application that uses a single camera and processes
information using a neural network specifically on the cloud.
5. CRITICAL DISCUSSION
In the last section, we discussed the different implementations of drowsy driving detection
systems and how each system was able to meet our two conditions of having a success rate of
over 80% and using minimal hardware. We also discussed how each implementation had
similarities and differences. At the end, each system met our conditions, but we wondered how
we could potentially make the systems work better. There have been many technical
advancements in the last couple of years that could potentially make these systems better to use.
The first tool we would recommend new researchers in the area would be to take advantage of
cloud computing. For example, the popularity of services such as Amazon Web Services (AWS),
Microsoft’s Azure, or Oracle’s cloud service are just a glimpse of what researchers could use to
create more effective machine learning processes without it bogging down the system on the
device itself. The main purpose of these services would be that instead of created a locally based
system and potentially hogging memory space, the algorithms that are necessary to detecting
drowsy driving would be computed on the cloud and sent to the device. This is key for mobile
implementations of drowsy driving detection. We need to take into consideration that the less
space that is necessary for the system can create better access for people who don’t have as
much storage on their mobile device or for creating hardware for less money. The drawback to
this system would be the need for a constant data connection so it wouldn’t be viable in countries
where cellular reception is not developed. The potential for creating a system that is integrated
with cloud computing can expand the use of the drowsy driving detection to other types of
detection system which could be more helpful as autonomous vehicles become more popular.
Another recommendation we would give to researchers would be to take advantage of the latest
cell phone hardware. As we were reading through our multiple sources, it became clear to us that
the papers we were reading were using dated hardware. With current technological
improvements, it should be mandatory to use the latest hardware. For example, a budget phone
with enough power would of cost around 200 to 250 US dollars. However, with the power of
competition, we can find hardware that has more processing power for even under 100 US
dollars. We can also expect that this advancement in hardware will continue to progress as
accordance to Moore’s law. These are a couple recommendations we can give that could help
further improve the current drowsy driving implementations.
6. PROS AND CONS
After looking through all our sources, we were able to identify the systems and how each one was
implemented. From our analysis, we were able to identify that many systems used the PERCLOS
metric to determine if the driver was drowsy or not. However, each system differed on how it
would use the PERCLOS metric. Some implementations would feed the eye region data into a
classifier to train its neural network so that it could predict if the driver was drowsy or not. Others
would measure different landmarks on the eyes to find the distance between these landmarks in
a measurement called EYA and would base their judgment system on that information. In this
section, we want to go over some of the pros and cons of each of these different systems
because even if some systems got close to 100% success rate, none could perfectly identify if the
driver was drowsy or not. From a general point of view, each of the systems that we discussed all
had a success rate of over 80%. The success rate is very important because it does not matter
how small the hardware is if the system cannot accurately detect if the driver is drowsy or not.
Another benefit to all these systems is that most of the hardware can fit easily into a vehicle. One
big drawback we identified is that some of these systems require fast processing to acquire data
and analyze in real-time, which means implementations of a phone’s processing power is difficult.
The systems that have these drawbacks are the systems that use machine learning or neural
networks, which trains a model to see if a driver would be falling asleep on the wheel or not.
These systems are meant to run for an extended duration where they are constantly working
profusely to detect and track eye movements. This amount of processing can be very demanding
on a system and requires a high-end computer. However, in research that utilized android
6. Samuel Kim, Irfan Wisanggeni, Ryan Ros & Rania Hussein
International Journal of Image Processing (IJIP), Volume (14) : Issue (1) : 2020 6
application and the PERCLOS method, drowsy driving detection was much better than the
machine learning and neural network systems. Unlike the other systems that matched faces to a
trained model, this one detects the face and monitors it by drawing boxes and lines to detect the
key facial landmarks, the eyes. An advantage that this implementation has is that the system was
able to run on much lower-end hardware and it utilized significantly less memory, which means it
can be implemented into an Android phone that has a working camera.
6. PREDICTED/EMERGING TRENDS
A trend that reoccurs in each research is the development of a system that is non-intrusive and
has high efficiency in terms of processing and algorithms. Most research projects on the topic of
drowsy driving detection have successfully identified working procedures, methods, and
implementations; In fact, the only problem that is left to solve is the size of the system and how to
implement it into a device that is non-intrusive. As a matter of fact, a research study has been
working to develop an algorithm that fits an embedded system by reducing the model size of a
neural network. By reducing the maximum memory size, the system takes less allocation of
memory when used and allows the system to be implemented into smaller devices, such as
Android smartphones. Another trend that is emerging is the accuracy of these systems to detect
drowsy drivers. Most of the research uses some form of a facial detection system, such as
tracking the eyes for behavioral studies, head movement, and yawning, while other systems
utilize machine learning algorithms to train a model through facial expressions in a database.
These methods have a success rating above 80%. Because other research methods are
intrusive, such as the implementation of EEG (electroencephalogram) and ECG
(electrocardiogram), newer research is leaning on the use of facial detection through computing
and algorithms since they are more accurate.
7. CONCLUSION
After analyzing multiple implementations, it became clear that the PERCLOS metric is highly
effective. Methods covered in the papers used different approaches to detect fatigue driving but
they all measured the success rate of their methods against the PERCLOS metric. The
effectiveness of PERCLOS presents a sufficient evidence that it is a reliable metric to be used by
researchers.
8. REFERENCES
[1] L. Tijerina, M. Gleckler, D. Stoltzfus, S. Johnston, MJ. Goodman and WW. Wierwille.
(1998, Sept.) “A Preliminary Assessment of Algorithms for Drowsy and Inattentive Driver
Detection on the Road,” Transportation Research Board. [On-line] Available:
https://ntlrepository.blob.core.windows.net/lib/17000/17900/17991/PB2001105783.pdf
[Feb. 12, 2020].
[2] Q. Ji, Z. Zhu, and P. Lan. (2004, Jul.) “Real-Time Nonintrusive Monitoring and Prediction of
Driver Fatigue,” IEEE Transactions on Vehicular Technology. [On-line] vol. 53, no. 4, pp.
1052–1068. Available: https://ieeexplore-ieee-
org.offcampus.lib.washington.edu/document/1317209 [Oct. 25, 2019].
[3] R. Grace, V. Byrne, D. Bierman, J.-M. Legrand, D. Gricourt, B. Davis, J. Staszewski, and B.
Carnahan. (2002, Aug.) “A drowsy driver detection system for heavy vehicles,” 17th DASC.
AIAA/IEEE/SAE. [On-line] Digital Avionics Systems Conference. Proceedings (Cat.
No.98CH36267). Available: https://ieeexplore-ieee-
org.offcampus.lib.washington.edu/document/739878 [Oct. 25, 2019]
[4] Liling Li, Mei Xie, Huazhi Dong. (2011, Sep.) “A method of driving fatigue detection based on
eye location” 2011 IEEE 3rd International Conference on Communication Software and
Networks. [On-line] Available IEEE Xplore Digital Library, https://ieeexplore-ieee
org.offcampus.lib.washington.edu/document/6013949 [Oct. 25, 2019]
7. Samuel Kim, Irfan Wisanggeni, Ryan Ros & Rania Hussein
International Journal of Image Processing (IJIP), Volume (14) : Issue (1) : 2020 7
[5] Mohsen Poursadeghiyan, Adel Mazloumi, Gebraeil Nasl Saraji, Mohammad Mehdi Baneshi,
Alireza Khammar, Mohammad Hossein Ebrahimi. (2018, Sep.) “Using Image Processing in
the Proposed Drowsiness Detection System Design” Available Iranian Journal of Public
Health. [On-line] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6174048/ [Oct. 25, 2019]
[6] Cui Xu, Ying Zheng, Zengfu Wang. (2008, Aug.) “Efficient eye states detection in real-time
for drowsy driving monitoring system” 2008 International Conference on Information and
Automation. [On-line] Available: IEEE Xplore Digital Library, https://ieeexplore-ieee-
org.offcampus.lib.washington.edu/document/4607990 [Oct. 25, 2019]
[7] Lunbo Xu, Shunyang Li, Kaigui Bian, Tong Zhao, Wei Yan. (2014, Apr.) “Sober-Drive: A
smartphone-assisted drowsy driving detection system” 2014 International Conference on
Computing, Networking and Communications (ICNC). [On-line] Available IEEE Xplore
Digital Library, https://ieeexplore-ieee-org.offcampus.lib.washington.edu/document/6785367
[Oct. 25, 2019]
[8] Ilkwon Park, Jung-Ho Ahn, Hyeran Byun. (2006, Sep.) “Efficient Measurement of Eye
Blinking under Various Illumination Conditions for Drowsiness Detection Systems” 18th
International Conference on Pattern Recognition (ICPR'06). [On-line] Available: IEEE
Xplore Digital Library, https://ieeexplore-ieee-
org.offcampus.lib.washington.edu/document/1698913 [Oct. 25, 2019]
[9] Cheung, Y., & Peng, Q. (2017). Method and Apparatus for Eye Gaze Tracking.