Steering wheel angle prediction by using dashcam images
Dataset was taken from here. Dataset contains 45,500 images taken from dashcam of the car along with angles.
A CNN model is used for prediction of the angles. L2 Loss function is used along with Adam optmizer and atan (inverse of tan) as activation function.
Basic flowchart of the complete process is given below-
Firstly, run train.py to train the model and save the trained model.
Use run_dataset.py to get live results as shown below.
Screenshot of the final application is shown below.
