Differential-UMamba: Rethinking Tumor Segmentation Under Limited Data Scenarios

2025-07-24

AI Summary

To address overfitting and poor generalization in few-shot medical image segmentation, this paper proposes Diff-UMamba—a novel architecture integrating the UNet encoder-decoder framework with Mamba’s long-range modeling capability. Its core contribution is the Signal Differential Noise Reduction Module (NRM), which explicitly suppresses noisy activations in the encoder via differential operations, thereby enhancing focus on lesion regions and improving task-relevant feature representation. This design significantly boosts feature learning robustness under low-data regimes. Experiments demonstrate that Diff-UMamba outperforms baseline models by 1–3% on public benchmarks including MSD, AIIB23, and BraTS-21, and achieves a 4–5% accuracy gain on the internal NSCLC dataset for gross tumor volume (GTV) segmentation. These results validate its effectiveness and superior generalization capability in few-shot medical image segmentation.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

In data-scarce scenarios, deep learning models often overfit to noise and irrelevant patterns, which limits their ability to generalize to unseen samples. To address these challenges in medical image segmentation, we introduce Diff-UMamba, a novel architecture that combines the UNet framework with the mamba mechanism for modeling long-range dependencies. At the heart of Diff-UMamba is a Noise Reduction Module (NRM), which employs a signal differencing strategy to suppress noisy or irrelevant activations within the encoder. This encourages the model to filter out spurious features and enhance task-relevant representations, thereby improving its focus on clinically meaningful regions. As a result, the architecture achieves improved segmentation accuracy and robustness, particularly in low-data settings. Diff-UMamba is evaluated on multiple public datasets, including MSD (lung and pancreas) and AIIB23, demonstrating consistent performance gains of 1-3% over baseline methods across diverse segmentation tasks. To further assess performance under limited-data conditions, additional experiments are conducted on the BraTS-21 dataset by varying the proportion of available training samples. The approach is also validated on a small internal non-small cell lung cancer (NSCLC) dataset for gross tumor volume (GTV) segmentation in cone beam CT (CBCT), where it achieves a 4-5% improvement over the baseline.

Problem

• Overcoming overfitting in tumor segmentation with limited data
• Enhancing focus on clinically meaningful regions via noise reduction
• Improving segmentation accuracy in low-data medical imaging scenarios

Innovation

• Combines UNet with Mamba for long-range dependencies
• Uses Noise Reduction Module to suppress irrelevant activations
• Improves segmentation accuracy in low-data settings