Synopsis
Breast cancer is one of the most common cancers worldwide, and early detection significantly improves survival rates. Mammography and histopathology images are widely used for breast cancer screening and diagnosis. However, manual analysis of these images by radiologists or pathologists can be time-consuming and prone to human error. We aim to develop a deep learning-based system for automated breast cancer detection using mammography and histopathology images. The model will leverage advanced architectures like Convolutional Neural Networks (CNNs) or Vision Transformers (ViTs) to classify images as benign or malignant and localize suspicious regions. This approach will assist in early diagnosis, reduce false positives/negatives, and support clinical decision-making.
Relevance of the Topic
Early and accurate detection of breast cancer is critical for improving patient outcomes. Automated systems can enhance diagnostic accuracy, reduce the workload on healthcare professionals, and provide consistent results across different imaging modalities. With the increasing availability of large annotated datasets and advancements in deep learning, there is significant potential to develop robust models that can aid in breast cancer screening and diagnosis.
Future Research/Scope
- Multi-modal Fusion: Combine mammography with other imaging modalities (e.g., ultrasound, MRI) to improve diagnostic accuracy.
- Explainability: Develop explainable AI techniques to highlight regions of interest in the images, ensuring transparency for clinicians.
- Real-time Screening: Optimize the model for real-time inference to integrate it into clinical workflows for faster diagnosis.
- Risk Stratification: Extend the model to predict the risk of breast cancer recurrence or progression based on image features.
- Generalization Across Populations: Investigate the model's performance across diverse populations to address biases in training data.
Skills Learned
- Deep Learning: Hands-on experience with CNNs, transfer learning, and Vision Transformers (ViTs).
- Medical Image Analysis: Understanding of mammography and histopathology image preprocessing, including normalization and augmentation.
- Python Programming: Proficiency in Python and libraries like TensorFlow, PyTorch, and OpenCV.
- Data Visualization: Skills in visualizing image data and model predictions using tools like Matplotlib and Plotly.
Relevant courses to the topic
- CSE428: Image Processing
- CSE422: Artificial Intelligence
- CSE425: Neural Networks
- CSE427: Machine Learning
- CSE443: Bioinformatics I
- Coursera: AI for Medical Diagnosis
Reading List
- Books
- "Deep Learning for Medical Image Analysis" – Zhou, Greenspan, Shen (Link)
- "Deep Learning for Biomedical Data Analysis" – Mourad Elloumi (Link)
- Research Papers
- "Deep Learning Based Methods for Breast Cancer Diagnosis: A Systematic Review and Future Direction" – Nasser et al., Diagnostics
Link - "Deep learning empowered breast cancer diagnosis: Advancements in detection and classification" – Jawad Ahmad et al., PLOS One
Link - "The Role of Deep Learning in Advancing Breast Cancer Detection Using Different Imaging Modalities: A Systematic Review" – Madani et al., Cancers
Link - "Deep learning algorithms for the early detection of breast cancer: A comparative study with traditional machine learning" – Martinez et al., Informatics in Medicine Unlocked
Link - "Application of Deep Learning in Breast Cancer Imaging" – Balkenende et al., Seminars in Nuclear Medicine
Link - "Strategies for Enhancing the Multi-Stage Classification Performances of HER2 Breast Cancer from Hematoxylin and Eosin Images" – Shovon et al., Diagnostics
Link - "Integrative hybrid deep learning for enhanced breast cancer diagnosis: leveraging the Wisconsin Breast Cancer Database and the CBIS-DDSM dataset" – Chandra Murty et al., Scientific Reports
Link - "Mammography with deep learning for breast cancer detection" – Lulu Wang, Frontiers in Oncology
Link - "U-Net: Convolutional Networks for Biomedical Image Segmentation" – Ronneberger et al., MICCAI
Link - "Review of Graph Neural Networks for Medical Image" – Jing Wang, EAI Endorsed Transactions on e-Learning
Link
- Datasets
- CBIS-DDSM (Curated Breast Imaging Subset of DDSM)
Link - Breast Cancer Histopathological Database (BreakHis)
Link - Invasive Ductal Carcinoma (IDC) Dataset [Widely Used, Binary Classification]
Link - Breast Cancer Digital Repository (BCDR)
Link - BACH (Breast Cancer Histology Images Challenge)
Link - MIAS (Mammographic Image Analysis Society Database)
Link - PCam (PatchCamelyon) [Simple Classification]
Link - TCGA-BRCA (The Cancer Genome Atlas - Breast Invasive Carcinoma) [For Multi-Modal]
Link - Breast Cancer Wisconsin (Diagnostic) [Tabular Data with Features Extracted from Images]
Link
- Code Tutorials & Repositories
- Videos & Playlists
