Analysis of Invasive Breast Cancer in Mammograms Using YOLO, Explainability, and Domain Adaptation

1. 1Implement YOLO architecture for object detection in mammograms
2. 2Use ResNet50-based OOD filtering to mitigate out-of-domain issues
3. 3Apply domain adaptation techniques to improve model reliability on varying equipment and imaging modalities
4. 4Evaluate model performance using explainability methods to provide insights into decision-making

Data scientists and AI engineers on a healthcare team can benefit from this research to improve the accuracy of breast cancer detection models, while radiologists and clinicians can use the explainability techniques to understand the model's decisions

💡 Domain adaptation and OOD filtering can significantly improve the reliability of deep learning models for breast cancer detection

💡 Improving breast cancer detection in mammograms with YOLO, explainability, and domain adaptation