Segmentation and Classification of Mammographic Images

This is our workspace during the experimental research process for the topic 'Deep Learning Techniques for Segmenting Breast Lesion Regions and Classifying Mammography Images.'

Note: We process data on a local machine and train on Kaggle using free GPUs.

Workflow

• We conducted data processing for the CIBS-DDSM dataset in the files main_train_cibs_ddsm and main_test_cibs_ddsm, and for the DDSM dataset in the file processing_data.
• The Train model folder contains files used for model training.
• We had evaluation alls model by model-evaluation.ipynb

Kaggle Links

• Repository for saving trained weights: https://www.kaggle.com/datasets/davidnguyens12/result/settings
• Model training notebooks:
  • https://www.kaggle.com/code/davidnguyens12/notebook-standard/script
  • https://www.kaggle.com/code/davidnguyens12/notebook-standard-apply-focal-loss/notebook
  • https://www.kaggle.com/code/davidnguyenss1/model-evaluation/notebook
  • https://www.kaggle.com/code/davidnguyens16/fork-of-sam-segmentation/notebook
  • https://www.kaggle.com/code/davidnguyens12/yolox