DETECT DRIVER Drowsiness and Distraction While Driving

Introduction

The project is the final project of the Computer Vision subject of group 15. The project focuses on developing a system to detect drowsiness and distraction while driving by using three different deep learning models: YOLO, CNN (Convolutional Neural Network) and ViT (Vision Transformer). When detecting a drowsy driver, the system will warn by emitting a sound.

Members:

• Nguyen Khac Luat – 21099741
• Hoang Ngoc Tan - 21074741
• Nguyen Thi Ty Ty - 21096511

File structure

• training_code/cnn-cv-01.ipynb: CNN model training notebook
• training_code/vit-01.ipynb: Vision Transformer model training notebook
• training_code/CodeTrainYolo.ipynb: YOLO model training with data
• link.txt: contains training notebooks and necessary datasets related to the project
• Report_Group15_ComputerVision.pdf: Detailed report about the project
• Slide_Group15_ComputerVision.pdf: Presentation slide program

You can use the dataset to train YOLO via the labeled API from Roboflow

from roboflow import Roboflow
rf = Roboflow(api_key="tOfbkzltAMQMayGo8I3p")
project = rf.workspace("luatluat").project("final2-kwf46-fkead")
version = project.version(2)
dataset = version.download("yolov11")

Kaggle Link

Model training notebooks can be found at:

• Vision Transformer (ViT): notebook
• CNN: notebook
• YOLO: notebook

Objectives

• Detect driver drowsiness
• Detect distracted driving behavior
• Timely warning to ensure traffic safety

Technology used

• Python
• Deep Learning
• Computer Vision
• CNN (Convolutional Neural Network)
• ViT (Vision Transformer)
• YOLO (ultralytics)

Installation and running instructions

System requirements

• Python 3.8 or higher
• GPU compatible with CUDA (recommended for model training)

Environment setup

1. Clone repository:

git clone https://github.com/khacluat03/final-project-computer-vision.git
cd final-project-computer-vision

2. Create a virtual environment (recommended):

python -m venv venv
# Windows
venv\Scripts\activate
# Linux/Mac
source venv/bin/activate

3. Install the required libraries:

pip install -r requirements.txt

Running the notebook

1. Open Jupyter Notebook:

jupyter notebook

2. Select and run one of the notebooks:

• cnn-cv-01.ipynb for CNN model
• vit-01.ipynb for Vision Transformer model
• YOLO for pre-trained YOLO model

Note

• Make sure all data is loaded into the data224x224/ directory
• Check that the GPU is installed and running correct
• The parameters in the notebook may need to be adjusted depending on the computer configuration.

References:

[1]. Khandave, A. (2020, September 20). Driver drowsiness detection alert system with Open-CV & Keras using IP-WebCam for camera connection. Retrieved from [https://www.linkedin.com/pulse/driver-drowsiness-detection-alert-system-open-cv-keras-khandave/]

[2]. Sahayadhas, A., Sundaraj, K., & Murugappan, M. (2012). Detecting driver drowsiness based on sensors: A review. Sensors, 12(12), 16937–16953. https://doi.org/10.3390/s121216937

[3]. Díaz-Santos, S., Cigala-Álvarez, Ó., Gonzalez-Sosa, E., Caballero-Gil, P., & Caballero-Gil, C. (2024). Driver identification and detection of drowsiness while driving. Applied Sciences, 14(6), 2603. https://www.mdpi.com/2076-3417/14/6/2603

[4]. Alshaqaqi, B., Baquhaizel, A. S., Ouis, M. E. A., Boumehed, M., Ouamri, A., & Keche, M. (2013). Driver drowsiness detection system. Trong 2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA). IEEE.

[5]. Liu, C. C., Hosking, S. G., & Lenné, M. G. (2009). Predicting driver drowsiness using vehicle measures: Recent insights and future challenges. Journal of Sleep Research, 18(3), 239-253. https://doi.org/10.1016/j.jsr.2009.04.005

[6]. Vural, E., Cetin, M., Ercil, A., Littlewort, G., Bartlett, M., & Movellan, J. (n.d.). Drowsy Driver Detection Through Facial Movement Analysis. Sabanci University; University of California San Diego. [7]. Grace, R., Byrne, V. E., Bierman, D. M., Legrand, J.-M., Gricourt, D., & Davis, B. K. (1998). A drowsy driver detection system for heavy vehicles. In 17th DASC. AIAA/IEEE/SAE. Digital Avionics Systems Conference. Proceedings (Cat. No.98CH36267). IEEE.

[8]. Zhang, H., Liu, T., Lyu, J., & Chen, D. (2023). Integrate memory mechanism in multi-granularity deep framework for driver drowsiness detection. Intelligence & Robotics, 3(4), 614-631. https://doi.org/10.20517/ir.2023.34