Segmenting a scene using the Fully Convolutional Network (FCN) model