This repository hosts the source code and related materials for the paper titled "Explainability in Deep Learning Segmentation Models for Breast Cancer by Analogy with Texture Analysis". Our project aims to advance the explainability of deep learning models in medical imaging, specifically in the context of breast cancer segmentation. By drawing analogies with texture analysis, we propose a novel approach to interpret the model's decisions, making these models more transparent and trustworthy for medical practitioners.

Before setting up the project, ensure you have the following installed on your system:

• Git
• Anaconda

conda create --name xai python=3.8
conda activate xai

The version used for the experiments for the paper is given below:

pip install torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 --index-url https://download.pytorch.org/whl/cu118

pip install torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2

pip install -U openmim
mim install mmengine==0.10.3
mim install mmcv==2.1.0

pip install mmsegmentation==1.2.2  

pip install ftfy regex mahotas scikit-image scikit-learn opencv-python pathlib segmentation-models-pytorch

git clone https://github.com/xrai-lib/xai-texture.git
cd xai-texture
cd src
python main.py

After completing the installation steps, you are ready to run the application. The main.py script is configured to demonstrate our methodology's application to breast cancer segmentation and explainability. The user interface will allow you to make use of all the functionalities available in the repository.

The original dataset used in this project is the Curated Breast Imaging Subset of DDSM (CBIS-DDSM), as described in the following paper:

• Paper: Lee RS, Gimenez F, Hoogi A, Miyake KK, Gorovoy M, Rubin DL. "A curated mammography data set for use in computer-aided detection and diagnosis research." Sci Data. 2017 Dec 19;4:170177.
• DOI: 10.1038/sdata.2017.177
• PMID: 29257132
• PMCID: PMC5735920

MMSegmentation Contributors (2020) MMSegmentation: OpenMMLab Semantic Segmentation Toolbox and Benchmark. https://github.com/open-mmlab/mmsegmentation