Multitask computer aided diagnosis of breast masses

In this page, a multitask computer aided diagnosis (CADx) framework is devised to investigate interpretability for classifying breast masses. We investigate interpretability in CADx with the proposed multitask CADx framework.

Dependencies

• MATLAB2014
• tensorflow >= 0.10.0rc0
• h5py 2.3.1
• scipy 0.18.0
• Pillow 3.1.0
• opencv2

Pretrained Models

Thanks to Davi Frossard you can download a pretrained weight here https://www.cs.toronto.edu/~frossard/vgg16/vgg16_weights.npz.

How to use

First, go into the 'data' directory and download "DDSM dataset"

For data consistency, mammograms scanned by Howtek 960 were selected from the DDSM dataset. By using LJPEG2PNG tool "DDSMUtility", load breast image. Preprocess the DDSM dataset to create a h5 file. It resizes images to 64*64.

After that preprecess the pretrained model vgg16_weights.npz

cd preprocess
python prepare.py

Finally, go into the 'code' directory and run 'main_ICADx.py'.

cd ..
python main_ICADx.py

You can see the samples in 'img and test' directory.

Reference Code

https://github.com/trane293/DDSMUtility