ExiST:面向半监督多器官分割的存在感知自训练方法
ExiST: Existence-Aware Self-Training for Semi-Supervised Multi-Organ Segmentation
AI 导读
论文提出存在感知自训练方法 ExiST,解决半监督多器官分割中自训练误将小器官推向消失的问题。该方法用独立训练网络的留一法共识构建伪标签,并以验证集质量门控筛选,对未确认器官采用集合值目标与边缘交叉熵训练。ExiST 在 Synapse 上以 20% 标注扫描取得 77.1 平均 Dice,在 AMOS 上以 5% 标注取得 75.9,超越现有 SOTA 方法。
正文 · 原文
Abstract:Semi-supervised multi-organ segmentation should learn more than a dozen organs from a handful of labeled scans, and the smallest organs occupy less than one percent of the foreground. We identify a failure mode of self-training that is specific to this regime. When a teacher misses a small organ in a scan, every voxel of that organ receives the confident pseudo-label background, and the student is trained toward the organ absence. Repeated over many scans, this loop extinguishes tail organs. In a pilot study, removing the gallbladder from the pseudo-labels reduced its Dice on AMOS from 36.5 to 1.0. We argue that a missing organ should be treated as unresolved rather than absent and propose Existence-aware Self-Training (ExiST). It builds pseudo-labels by leave-one-out consensus of independently trained networks, screens them with a quality gate on validation scans, and infers, per scan and per organ, whether the organ was found. For every unresolved organ, the background voxels of that scan receive the set-valued target, trained with a marginal cross-entropy that accepts either answer and a Dice term that skips uncertain voxels. We show that this loss never pushes a candidate organ toward absence and never favors one member of the set, which leaves the decision to labeled evidence. ExiST outperforms the current state-of-the-art methods and reaches 77.1 mean Dice on Synapse with 20% labeled scans and 75.9 on AMOS with 5% labeled scans.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.23581 [cs.CV] |
| (or arXiv:2611.23581v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2611.23581 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Maxx Richard Rahman [view email]
[v1]
Tue, 17 Nov 2026 22:27:46 UTC (3,812 KB)
来源:arXiv cs.AI · arxiv.org