This project is NOT for commercial use. You are NOT allowed to share this project unless with our permission. Step 0: Intall Matlab CUDA CUDNN in Win10 System. Step 1: Put cell images in data/cells. Crop those images with size of 600*600 to 450*450. Step 2: Install vlfeat and matconvnet to your path of matlab toolbox. Step 3: Download pre-trained models from http://www.vlfeat.org/matconvnet/pretrained/imagenet-resnet-50-dag.mat http://www.vlfeat.org/matconvnet/pretrained/imagenet-resnet-101-dag.mat http://www.vlfeat.org/matconvnet/pretrained/imagenet-resnet-152-dag.mat and put them in images in data/models. Step 4: Create a new file with name isbi_test_guess.csv and paste the follow content in it. It contains 'label' and 2586 binary numbers. label 1 0 0 1 1 .... Step 5: Run Phrase_III_of_ISBI_NMC.m Step 6: You can also download our trained basic models from https://drive.google.com/drive/folders/1DPc85NsP4PiQz39x3up55Qx7KoP0iw9g?usp=sharing https://drive.google.com/drive/folders/177dyrKsWdkP5XoNhsaD6Cp1wyIdx3aDt?usp=sharing