1. Real Time Driver Drowsiness Detection
Using Haar Cascade Algorithm
Presented by
SANDHIYA V
SARANYA C
SARANYA S
Under the guidance of
Mr. Murudass, MSC. M.Phil., MCA.,
2. ABSTRACT
To check the driver fatigue by detecting the facial expressions of
the drivers.
Using face detection technique.
On the Dlib toolkit, the lankmark of frontal driver facial emotion
in a frame are found.
By taking the eyes aspect ratio as input.
If the drivers drowsiness detection it generates voice alert to
make the driver awake .
3. System specification
• Hardware Specification
• Processor : INTEL I5 (7th generation)
• RAM : 4 GB RAM
• Hard disk : 1TB
• Monitor : 20’ color monitor
•
• Software Specification
• Front end : PyQt5 designer
• Back end : python
• Software tool used : Visual code studio
• Platform : Windows 8
4. EXISTING SYSTEM
In pre-existing features for facial landmark detection
It is implemented to identify the state of drowsiness and fatigue.
68- facial landmark predefined landmark helps in shape prediction
It identify the various regions of the face like
Eye brows,
Eye,
Mouth region,
etc.,
5. DISADVANTAGES
The algorithm used here, rely on the detected facial landmarks.
It make the system less sensitive to landmark misalignment.
It Cannot detect exact drowsiness.
6. PROPOSED SYSTEM
It is being developed is focused on image recognition based on drowsiness.
Drowsiness can be detected by scrutinizing the status of eye.
This algorithm which will automatically extract eye area.
And Scrutinize this eye part for drowsiness detection based on status of
eye.
It is framework is done on the based of real time face.
It gives a voice alert to driver when drowsiness is detected.
7. ADVANTAGES
It detecting the drowsiness of the driver,
The system also alerts the driver to make him awake at the right time.
It achieves high precision in detection of drowsiness.
8. MODULES
There are five modules in this system:
• Real-time camera monitoring module.
• Face region detection module.
• Eyes part scrutinization module .
• Drowsiness detection module.
• Voice alert module.
9. REAL TIME CAMERA MONITORING
The real-time camera is used to get input image.
As the camera will capture live video streaming,
Convert the video into image frames
By extracting the images from video.
10. FACE REGION DETECTION
It used HAAR cascade classifier to detect.
It detect face region in the images.
11. EYES PART SCRUTINIZATION
The Dlib toolkit is used to extract prime facial features.
It present in the face region of the image.
Facial landmark algorithm will efficiently locate the region of interest
For Exemple
• Eye Region
12. DROWSINESS DETECTION
The drowsiness of the driver will be detected.
By determining the eyes aspect ratio.
Eyes aspect ratio is calculated.
By estimating the changes in that parameter’s value.
It can determine the drowsy state of the driver.
13. VOICE ALERT
The driver will be warned by generating voice alert
The driver is detected to be in drowsy state
To avoid those serious consequences,
The system alerts the driver to make him attentive right after the detection
of drowsiness.
14. Conclusion
HAAR cascade classifier and facial landmark algorithm is used to detect
many faces and eyes, nose, mouth etc.
It detection of features increases rapidly when a camera of high
specifications is used.
We are implementing this system in vehicles,
It only the driver’s facial expression.
The face is detected and recognized as drowsy,
It will get an voice alert to driver.