Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Beyond Random Sampling: Distribution-Aware Alignment for Semi-Supervised Medical Image Segmentation

About

Precise medical image segmentation is crucial for clinical diagnosis and treatment planning, yet relies heavily on expensive expert annotations. Semi-supervised medical image segmentation (SSMIS) offers a cost-effective solution but typically operates under the assumption of independent and identically distributed (i.i.d.) data, defaulting to random sampling. While statistically valid at scale, this strategy suffers from severe representation bias in low-data regimes, failing to capture the heterogeneous medical data manifold. To address this, we propose a highly data-efficient framework driven by distribution alignment. First, we introduce an offline Distribution-Aware Sample Selection strategy. By leveraging Vision Foundation Models (VFMs) and our designed Density-K-Center algorithm, we explicitly identify representative structural anchors, establishing a more representative labeled domain. Second, to bridge the remaining distribution gap, we propose the Memory-guided Copy-Paste (MCP) module. Tailored for the inherent class imbalance in medical scans, MCP leverages a semantic memory mechanism to retrieve historically consistent priors for cross-domain alignment, encouraging semantic consistency. Coupled with an easy-to-hard progressive schedule, this framework effectively mitigates early-stage pseudo-label noise. Extensive experiments on six diverse 2D and 3D datasets demonstrate strong segmentation performance, particularly in extremely low-labeled scenarios (\eg, 1/16 ratio).

Weihao Yan, Yeqiang Qian, Yi Dong, Ming Yang• 2026

Related benchmarks

TaskDatasetResultRank
Medical Image SegmentationISIC
DICE89
114
2D Medical Image SegmentationBUSI
IoU75.8
41
Medical Image SegmentationCAMUS
Overall Dice Score93.2
31
2D Medical Image SegmentationPMTCXR
IoU41.6
27
Medical Image SegmentationACDC (10%)
Dice91
23
3D Medical Image SegmentationPROMISE Low Ratio 1-16 v1
Mean IoU77.7
9
3D Medical Image SegmentationPROMISE Medium Ratio (1/10) v1
IoU78.4
9
3D Medical Image SegmentationPROMISE High Ratio (1/5) v1
IoU77.5
9
3D Medical Image SegmentationACDC Low Ratio (1/20) v1
IoU82.8
9
3D Medical Image SegmentationACDC High Ratio (1/5) v1
IoU84.6
9
Showing 10 of 10 rows

Other info

Follow for update