DSU-Net: An Attention-Enhanced Dense Skip U-Net for Breast Lesion Segmentation in Mammographic Images

2026-06-08 · Source: cs.CV updates on arXiv.org · Field: Science & Research — Health & Medical Research, Mathematics & Computational Sciences · Depth: Expert, quick

Summary

DSU-Net, an attention-enhanced Dense Skip U-Net architecture, is presented for automated breast lesion segmentation in mammographic images. This framework integrates dense skip connections and attention mechanisms to improve feature propagation, preserve spatial information, and enhance lesion boundary delineation. Experiments utilized the Curated Breast Imaging Subset of the Digital Database for Screening Mammography (CBIS-DDSM) and addressed foreground-background imbalance with a composite loss function combining Dice loss, focal loss, and binary cross-entropy loss. The model achieved a Dice Similarity Coefficient of 0.9421, an Intersection over Union of 0.8905, an accuracy of 0.9711, and an AUC-ROC of 0.9878 on the validation dataset. These results demonstrate DSU-Net's accurate and reliable breast lesion segmentation, highlighting attention-guided deep learning's potential for computer-aided breast cancer screening.

Key takeaway

For AI Scientists developing medical imaging solutions, DSU-Net's performance suggests integrating attention mechanisms and dense skip connections significantly improves segmentation accuracy. You should consider composite loss functions, like Dice, focal, and binary cross-entropy, to effectively manage severe foreground-background imbalance in your datasets. This approach can lead to more robust and reliable computer-aided diagnostic tools.

Key insights

DSU-Net combines dense skip connections and attention for accurate breast lesion segmentation in mammography.

Principles

• Attention mechanisms enhance feature relevance.
• Dense skip connections improve information flow.
• Composite loss functions mitigate class imbalance.

Method

DSU-Net employs a U-Net architecture with dense skip connections and attention modules, trained on CBIS-DDSM using a composite loss of Dice, focal, and binary cross-entropy.

In practice

• Apply attention-guided deep learning for medical image analysis.
• Use composite loss for imbalanced segmentation tasks.

Topics

• Breast Lesion Segmentation
• Mammography
• DSU-Net
• U-Net Architecture
• Attention Mechanisms
• Deep Learning

Best for: Computer Vision Engineer, AI Scientist, Research Scientist

Related on AIssential

• Attention-Gated Convolutional Networks for Scanner-Agnostic Quality Assessment — A hybrid CNN-Attention model effectively assesses MRI quality, generalizing across diverse scanners without retraining.
• Improving breast cancer screening workflows with machine learning — Google Research, in partnership with NHS organizations, conducted the Artificial Intelligence in Mammography Screening (AIMS) study, publishing two companion studies in Nature Cancer to evaluate an AI…
• A 3D SAM-Based Progressive Prompting Framework for Multi-Task Segmentation of Radiotherapy-induced Normal Tissue Injuries in Limited-Data Settings — A 3D SAM-based framework improves radiotherapy injury segmentation using progressive prompting and a small-target focus loss.
• Margin-Consistent Deep Subtyping of Invasive Lung Adenocarcinoma via Perturbation Fidelity in Whole-Slide Image Analysis — Margin consistency with Perturbation Fidelity enhances deep learning robustness and accuracy for lung adenocarcinoma subtyping.
• A deep learning framework for glomeruli segmentation with boundary attention — A U-Net-based model with a boundary attention decoder improves glomeruli segmentation in kidney tissue.
• DenseNet Paper Walkthrough: All Connected — DenseNet uses aggressive skip connections to mitigate vanishing gradients and enhance information flow in deep neural networks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.CV updates on arXiv.org.