rohan2516/Distracted_Driver_CNN

Built a Convolutional Neural Network to Classify Drivers' Behaviour. The model classifies them into different categories such as texting, eating, talking on the phone, makeup, reaching behind, etc. Data from the images is extracted by using ImageDataGenerator. The model makes use of three Conv2D layers each followed by a MaxPool2D layer. To avoid overfitting each three Dropout layers are used. The model contains three Dense layers with a varying number of units to accurately predict the Drivers' Behaviours. The model is trained on 17,000 images and makes predictions on 5,600 images with an accuracy of 99.5%.

This repository is not active