Driver Drowsiness

Introduction

Implementation of driver drowsiness detection with EfficientNetB7 in TensorFlow Keras and OpenCV.

Installation

• Install the requirements: pip install -r requirements.txt

Training

• Arguments which you change as desired:
  • Image size
  • Training batch size
  • Maximum epochs
  • Test data ratio
• Haarcascade XMLs saved in folder prediction_images - For face: haarcascade_frontalface_default.xml - For eye: haarcascade_eye.xml

Training

Run the train.py to train the model with the below command:

• python .\train.py --Image_size 145 --Batch_size_train 20 --maximum_epochs 200 --Test_size_ratio 0.15

Prediction

• Run the predict.py to test the model saved as drowsiness_newB7.h5

You can change the image and the conditions to check for drowsiness as required. This is just a basic application with more updates coming in the future.

Loss and Accuracy Plots

Citation

@inproceedings{tan2019efficientnet, title={Efficientnet: Rethinking model scaling for convolutional neural networks}, author={Tan, Mingxing and Le, Quoc}, booktitle={International conference on machine learning}, pages={6105--6114}, year={2019}, organization={PMLR} }