In this project, the road in images is detected using a Fully Convolutional Network (FCN).
The functions in main.py are implemented to train a FCN.
The result of the FCN applied to the test dataset can be found runs.zip.
The training is performed on the Kitti Road dataset which can be downloaded from here.
The dataset needs to be extracted to the data folder in order to run the code. This will create the folder data_road with all the training a test images.
The following needs to be installed:
Run the following command to run the project:
python main.py
All images can be found in runs.zip.
