Final Project in Deep Learning

This is the final project in a deep learning course at Ben Gurion University. In this project we had implemented a number of Architectures for Semantic Segmentation of cells from microscope images.

Running the model

python mainScript.py --mode='train' --net-type='FCN' --plot-activations=true --plot-weights=false

the following command will run train a fully convolutional architecture with plots of the activation functions.

The net types are the following:

• FCN
• EncoderDecoder
• ResNetFCN
• DeeperResNetFCN
• EncoderDecoderResNet