Department of Computer Science and Engineering

Mini Project (BCS586)

Review – 0
DROWSINESS DETECTION FOR DRIVERS
Batch No.: 2025_CSE_08

AKSHAY KRSHNA 1KS22CS009

Guided by
ADITYA S 1KS22CS005
Mr.Abhilash L Bhat
DARSHAN D GOWDA 1KS22CS037
Assistant Professor
ABHISHEK A 1KS22CS002

1
 Contents

 Overview
 Goals
 Applications
 Requirement Specifications
 Control Flow
 Data Set
 References

2
 Overview
 Driver fatigue is a significant factor in a large number of vehicle accidents. Thus, driver
Drowsiness detection has been considered a major potential area so as to prevent a
huge number of sleep induced road accidents.
 The Driver Drowsiness Detection System aims to enhance road safety by detecting
drowsiness through eye and facial analysis.
 Using a webcam and software tools like Python, OpenCV, and Dlib, the system
captures video input, preprocesses frames (converting to grayscale and applying
Gaussian blur), and detects faces and eyes.
 By calculating the Eye Aspect Ratio (EAR) and setting a threshold, the system triggers
alerts when drowsiness is detected. Testing under various conditions ensures accuracy,
and documentation includes a report and presentation for project defense.

3
 Goals

• Enhance Road Safety: Prevent accidents

caused by driver fatigue.
• Real-Time Monitoring: Analyse eye states
using image processing techniques.
• Increase Detection Accuracy: Improve
precision through advanced algorithms.
• Alert Mechanism: Provide timely audio
or visual alerts for drowsiness.
• Research and Development: Explore IoT
and machine learning for enhanced
system capabilities.

4
 Applications
• Accident Prevention: Reduces the risk of accidents caused by drowsy driving by
providing timely alerts to drivers.
• Commercial Fleet Management: Enhances safety for long-haul truck drivers and
delivery services by monitoring driver alertness during extended hours.
• Smart Vehicle Integration: Integrates with advanced driver-assistance systems (ADAS)
to improve overall vehicle safety and performance.
• Insurance Benefits :Potentially lowers insurance premiums for vehicles equipped with
drowsiness detection systems due to reduced accident risk.
• Public Transportation Safety: Ensures the safety of passengers in buses and taxis by
monitoring driver fatigue levels.
• Research and Development: Provides data for further studies on driver behavior and
fatigue, contributing to advancements in transportation safety technologies

5
 Requirement Specifications
• Functional Requirements: Image Capture: Use a webcam or camera to capture driver images
continuously at 20fps.Face Detection: Implement algorithms to detect the driver's face .Eye
Detection: Identify eye regions and calculate the Eye Aspect Ratio (EAR) to assess
drowsiness. Alert Mechanism: Trigger audio or visual alerts when drowsiness is detected.
• Non-Functional Requirements: Reliability: Ensure accurate detection of drowsiness to
promote driver safety. Real-Time Response: Provide immediate alerts to avoid delays in
detection. Unobtrusive Operation: The system should be transparent and not distract the
driver. Economical: Design the system to be cost-effective for widespread implementation.
Flexibility: Accommodate various user physical attributes for effectiveness.
• Constraints: Power Consumption: Operate efficiently on limited power sources available in
vehicles.
• Technology Requirements: Utilize software tools such as Python, OpenCV, and Dlib for
image processing and analysis. 6
Control Flow

7
 Data Set
• YawDD Dataset: Videos from dashboard cameras featuring
diverse drivers. Used for training and evaluating drowsiness
detection systems.
• Closed Eyes in the Wild (CEW) Dataset: Contains images of
2,423 subjects with open and closed eyes. Sourced from the
Labeled Face in the Wild dataset.
• Drowsiness Detection Dataset (Roboflow): Collection of images
depicting "drowsy" and "awake" facial expressions in a vehicle.
• Driver Drowsiness Dataset (D3S):Includes videos and frames of
various states (eye closed, yawning, etc.) for training purposes.

8
 References
• Sun, Y., Wu, C., Zhang, H., Zhang, Y., Li, S., & Feng, H. (2021). Extraction of Optimal
Measurements for Drowsy Driving Detection Considering Driver Fingerprinting
Differences. Journal of Advanced Transportation, 1–17.
http://dx.doi.org/10.1155/2021/5546127.
• J. S., & Smitha P. S. (2014). Driver Drowsiness Detection System. IOSR Journal of VLSI
and Signal Processing, 4(1), 34–37. http://dx.doi.org/10.9790/4200-04113437.Gargano,
I. E. (2022).
• Model-Based Validation of Driver Drowsiness Detection System for ADAS. Master's
thesis, Alma Mater Studiorum - Università di Bologna. http://amslaurea.unibo.it/25716/.
• Svensson, U. (2004). Blink Behaviour Based Drowsiness Detection: Method
Development and Validation. Swedish National Road and Transport Research Institute.
http://www.vti.se.
• Abas, A. B. (2011). Non-Intrusive Driver Drowsiness Detection System. Thesis,
University of Bradford. http://hdl.handle.net/10454/5521.