BUCAN: Bayesian Uncertainty-aware Classification with Attention Networks for Medical Images
Sagar, A.
Show abstract
Accurate and reliable medical image classification is critical for clinical decision-making across diverse imaging modalities, including X-ray, CT, and MRI. Traditional convolutional neural networks often produce overconfident predictions, limiting their clinical trustworthiness. In this work, we propose an uncertainty-aware, attention-augmented neural network that integrates multi-scale SwirlAttention and FeedBackAttention modules with a Bayesian probabilistic classifier. This framework enables robust feature extraction, interpretable attention maps, and principled estimation of epistemic uncertainty. We evaluate our approach on four diverse datasets, including Diabetic Retinopathy, Kvasir, Skin Cancer, and fused multi-focal Oocyte images, covering a wide range of pathological and morphological variations. Extensive experiments demonstrate that our method outperforms state-of-the-art CNN and transformer-based baselines in terms of accuracy, calibration, and interpretability. Grad-CAM visualizations highlight clinically relevant regions, while uncertainty estimates provide actionable insights for ambiguous cases, making the framework suitable for reliable deployment in real-world clinical settings.
Matching journals
The top 11 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- STAMP: Simultaneous Training and Model Pruning for Low Data Regimes in Medical Image Segmentation 96%
- A Framework for Falsifiable Explanations of Machine Learning Models with an Application in Computational Pathology 95%
- Clinical Validation of Saliency Maps for Understanding Deep Neural Networks in Ophthalmology 94%
Similar papers in this journal
- SimSearch: A Human-in-the-Loop Learning Framework for Fast Detection of Regions of Interest in Microscopy Images 96%
- Dual-Field Microvascular Segmentation: Hemodynamically-Consistent Attention Learning for Retinal Vasculature Mapping 95%
- pathCLIP: Detection of Genes and Gene Relations from Biological Pathway Figures through Image-Text Contrastive Learning 92%
Similar papers in this journal
- BenchXAI: Comprehensive Benchmarking of Post-hoc Explainable AI Methods on Multi-Modal Biomedical Data 96%
- MultiHeadGAN: A Deep Learning Method for Low Contrast Retinal Pigment Epithelium Cells Segmentation in Fluorescent Flatmount Microscopy Images 96%
- Deep Multimodal Graph-Based Network for Survival Prediction from Highly Multiplexed Images and Patient Variables 95%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.