CCS355 Neural Network & Deep Learning UNIT III notes and Question bank .pdf
Final PPT.pptx
1. Project
on
“Real-time Face Detection & Security System for ATM
Machine”
Guided By : Prof. Bhagyashree Kadam
JSPM’s
Bhivarabai Sawant Institute of Technology & Research
Accredited with ‘B++’ Grade by NAAC
Presented By
Department of Information Technology
1. Akash Jadhav - 72171616E
2. Kishor Dukre - 72171613L
3. Vaibhav Kumbhar - 72171622K
4. Shivraj Sakunde - 72171626B
2. CONTENTS
– Introduction
– Problem statement
– Literature survey
– Existing System
– Proposed system
– System Architecture
– Algorithm
– Data Flow Diagram
– Component diagram
– Use case diagram
– Output
– System requirement
– Advantages and Limitations
– Conclusions and Future work.
– List of publications
– References
3. INTRODUCTION
Face recognition-based ATM system will revolutionized the way we interact with banking
services, combining advanced technology with convenience and enhanced security. With the
ability to authenticate users based on their unique facial features, this system will replace the
traditional reliance on passwords and PINs.
By leveraging sophisticated algorithms and high-resolution cameras, face recognition ATM offer a
seamless and user-friendly experience, ensuring rapid and accurate identification of customers.
This innovative approach not only streamlines banking operations but also mitigates the risks
associated with card fraud and identity theft.
Face recognition eliminates the need for physical ATM cards, making transactions more
convenient for users. Users no longer need to carry their cards, remember PINs, or worry about
losing their cards. They can simply approach an ATM, look into the camera, and the system will
authenticate their identity.
3
4. LITERATURE REVIEW
Sr. No. Paper Title Year Description
1
Autonomous Face Detection
System from Real-time Video
Streaming for Ensuring the
Intelligence Security System
IEEE (2020)
2020
Face detection is a popular research topic to the
researchers at present in the
Biometric system. The aim of this research is to
develop a system which can detect
human face or faces from a live video streaming
2
Systematic Review of
Intelligence Video
Surveillance: Trends,
Techniques, Frameworks,
and Datasets” IEEE (2019)
2019
In this paper, Intelligent Video Surveillance (IVIST)
is also introduced. IVIST is a prototype system with
EDCAR knowledge that provides automatic support
to implement related inference procedures such as
object tracking and detection to trigger alarms.
3
Authentic Face Detection and
Encryption for Security
Assurance IEEE (2018)
2018
The rapid evolution of authentic face detection has
been witnessed from the past
along with the security assurance in the field of
information technology. Now being a necessary note
of issue the concept of face detection along with
encryption.
5. PROBLEM STATEMENT
Use face recognition technology to develop an ATM system that will offer a user-
friendly and secure alternative for individuals who have difficulty remembering
passwords. provide a convenient and accessible way to access banking services,
reducing the stress and frustration associated with traditional password-based
authentication systems. Forgetting passwords can be frustrating, especially when it
comes to accessing important services like banking.
6. MOTIVATION
The face recognition technology has advanced significantly in recent years, allowing
for efficient recognition even in different lighting conditions, angles, and facial
expressions. With face recognition technology, individuals no longer need to
remember complex passwords or PINs to access their accounts. Instead, their facial
features serve as their unique identifier. Face recognition eliminates the need for
physical ATM cards, making transactions more convenient for users. Users no longer
need to carry their cards, remember PINs, or worry about losing their cards.
7. ABSTRACT
Our face recognition-based ATM system project aims to develop a secure and convenient
authentication method for accessing ATM services. Traditional ATM systems rely on physical cards
and PINs, which can be susceptible to theft, fraud, or forgotten passwords. In this project, face
recognition technology is leveraged to authenticate users based on their unique facial features,
eliminating the need for cards and PINs. The system utilizes high-resolution cameras and advanced
algorithms to capture and analyze facial images, comparing them against a database of enrolled
users. The project focuses on addressing challenges such as accuracy, privacy, accessibility, and
security. By implementing face recognition technology, the project aims to provide a seamless and
user-friendly experience for ATM users while enhancing security and mitigating the risks associated
with traditional authentication methods. Through rigorous testing and evaluation, the project aims to
demonstrate the feasibility and effectiveness of face recognition-based ATM systems in real-world
banking environments.
8. SYSTEM ARCHITECTURE
Human Live Face
Recognition
Webcam Capture Image Preprocess Image
Feature Extraction
(SIFT Key Points )
Classification
(based on SVM)
Face Recognition
(Authorize)
Open ATM
Training
Face Database
Unauthorized
No Login
9. ALGORITHM
Haar Cascade algorithm:
1. Collecting Positive and Negative Images: Collect a set of positive and negative images. Positive
images should contain the object you want to detect, and negative images should not contain the object.
2. Preparing the Training Data: Convert the positive and negative images to grayscale, resize them to
the same size, and store them in a file with labels indicating whether they are positive or negative.
3. Creating the Haar Features: The Haar features are created by defining rectangular regions in the
image and calculating the difference between the sum of pixel values inside the rectangle and the sum of
pixel values outside the rectangle.
4. Training the Classifier: Train the classifier using the positive and negative images and the Haar
features. This is done using a machine learning algorithm such as AdaBoost or Support Vector Machines
(SVM).
10. 5. Creating the Cascade Classifier: The trained classifier is converted into a cascade of
classifiers, which are applied to sub-regions of the image in a hierarchical manner. This
allows the algorithm to quickly reject regions that are unlikely to contain the object.
6. Detecting Objects in Images: The cascade classifier is applied to the image to detect
objects. This is done by moving a sliding window over the image and applying the cascade
classifier to each sub-region.
7. Filtering False Positives: The algorithm may detect false positives, which are regions
that are not actually the object of interest. These false positives are filtered out using various
techniques such as non-maximum suppression.
8. Fine-tuning: Finally, the algorithm is fine-tuned to improve its accuracy and reduce false
positives. This can be done by adjusting various parameters such as the scale of the sliding
window and the threshold for detecting the object.
19. ADVANTAGES
1.Password-Free Authentication: With face recognition technology, individuals no longer
need to remember complex passwords or PINs to access their accounts. Instead, their facial
features serve as their unique identifier.
2.Rapid and Accurate Authentication: Face recognition systems can authenticate
individuals quickly and accurately.
3.Increased Convenience: Face recognition eliminates the need for physical ATM cards,
making transactions more convenient for users. Users no longer need to carry their cards,
remember PINs, or worry about losing their cards..
4.Enhanced User Experience: Face recognition-based ATMs provide a more user-friendly
experience.
20. SYSTEM REQUIREMENTS
Software Requirements Minimum Recommended
OS Windows 7 Windows 10 or higher
Database SQLite 3.34 SQLite 3.42
IDE VS Code 1.52 VS Code 1.78
Python python 3.8 Python 3.11
Hardware Requirements Minimum Recommended
Processor Intel core i5 10th gen Intel core i7 12th gen
Camera 16 MP 48 MP
RAM 8 GB 16 GB
HDD 2GB 4 GB or more
21. CONCLUSION
In conclusion, a real-time person detection project has several potential applications in various
industries such as security, surveillance, retail, and healthcare. The project involves using
computer vision techniques to detect people in real-time video streams, and it can be achieved
through various methods such as Haar cascades, deep learning, and other machine learning
algorithms.
22. FUTURE SCOPE
1. Contactless Transactions: The COVID-19 pandemic has accelerated the demand for
contactless transactions. An ATM system using face detection technology eliminates the need
for physical touchpoints, providing a more hygienic option.
2. Personalization: The ATM system can personalize the user experience by recognizing their
face and displaying preferred language, currency, and transaction history.
3. Data Collection: The system can collect data on user preferences and transaction patterns,
enabling banks to offer targeted marketing and personalized promotions.
4. Enhanced Security: With face detection technology, the ATM system can identify the user
based on their facial features, making it more secure and reducing the chances of fraud or theft.
23. LIMITATIONS
1. False Positives: Real-time person detection projects can generate false positives, which can
result in unnecessary alarms or interventions.
2. Technical Limitations: The accuracy of real-time person detection projects can be affected by
factors such as lighting, environmental conditions, and technical limitations of the sensors or
algorithms used.
3. Biases and Discrimination: Facial recognition systems have been found to exhibit biases,
including racial, gender, and age biases. If not properly trained or calibrated, these systems may
disproportionately misidentify certain groups, leading to discrimination and exclusion.
24. REFERENCES
1. Face Recognition in Real-world Surveillance Videos with Deep Learning Method” Department of
Information and Technology University of Science and Technology of China IEEE (2017)
2. Authentic Face Detection and Encryption for Security Assurance IEEE (2018)
3. Face Detection based ATM Security System using Embedded Linux Platform IEEE (2017)