In this project, trained the Full Convolutional Network to identify road pixels in the image. Starting with the VGG16 pretrained model, Built FCN by adding a skip layer, 1x1 convolution and transposed convolution.
It is very easy. Just run main.py. using a GTX 1070 it takes about 10 minutes to train.
Make sure you have the following is installed:
Download the Kitti Road dataset from here. Extract the dataset in the data folder. This will create the folder data_road with all the training a test images.