Mammo-Clustering: A Weakly Supervised Multi-view Tri-level Information Fusion Context Clustering Network for Localization and Classification in Mammography

2024-09-23

AI Summary

To address the problem of missed detection of microcalcifications and subtle lesions (e.g., early-stage malignancies) in mammograms due to downsampling of high-resolution images, this paper proposes a weakly supervised, multi-view, three-level contextual clustering network (CMC-Net). Methodologically, CMC-Net introduces a novel hierarchical fusion mechanism—integrating global, feature-level local, and patch-level local representations—and is the first to incorporate a lightweight, structurally aware contextual clustering paradigm into breast cancer screening, thereby preserving fine-grained details while ensuring clinical interpretability. Evaluated on Vindr-Mammo and CBIS-DDSM, CMC-Net achieves AUCs of 0.828 and 0.805, respectively—surpassing the best prior baselines by 3.1% and 2.4% (p < 0.05). These results demonstrate statistically significant improvements in detecting subtle pathological findings.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Breast cancer is a significant global health issue, and the diagnosis of breast imaging has always been challenging. Mammography images typically have extremely high resolution, with lesions occupying only a very small area. Down-sampling in neural networks can easily lead to the loss of microcalcifications or subtle structures, making it difficult for traditional neural network architectures to address these issues. To tackle these challenges, we propose a Context Clustering Network with triple information fusion. Firstly, compared to CNNs or transformers, we find that Context clustering methods (1) are more computationally efficient and (2) can more easily associate structural or pathological features, making them suitable for the clinical tasks of mammography. Secondly, we propose a triple information fusion mechanism that integrates global information, feature-based local information, and patch-based local information. The proposed approach is rigorously evaluated on two public datasets, Vindr-Mammo and CBIS-DDSM, using five independent splits to ensure statistical robustness. Our method achieves an AUC of 0.828 on Vindr-Mammo and 0.805 on CBIS-DDSM, outperforming the next best method by 3.1% and 2.4%, respectively. These improvements are statistically significant (p<0.05), underscoring the benefits of Context Clustering Network with triple information fusion. Overall, our Context Clustering framework demonstrates strong potential as a scalable and cost-effective solution for large-scale mammography screening, enabling more efficient and accurate breast cancer detection. Access to our method is available at https://github.com/Sohyu1/Mammo_Clustering.

Problem

• Improves breast cancer detection in mammography images.
• Enhances localization and classification with triple information fusion.
• Increases computational efficiency and accuracy in clinical mammography tasks.

Innovation

• Context Clustering Network
• Triple information fusion
• Weakly supervised multi-view clustering