SCKAN: Structural Consensus-based KAN Prototype Learning for Semi-Supervised Pancreas Segmentation

2026-05-26

AI Summary

This work addresses the challenges of supervision bias and limited generalization in semi-supervised pancreatic segmentation, which arise from substantial morphological variations across subjects. To mitigate these issues, the authors propose a cross-sample structural consensus learning framework that introduces, for the first time, prototype consistency learning under structural constraints. The method integrates Kolmogorov–Arnold Networks (KANs) with B-spline adaptive nonlinear mechanisms to effectively alleviate morphology-specific biases under sparse annotation settings. By jointly optimizing prototype-level contrastive learning and enforcing structural consistency, the model achieves significantly superior performance compared to existing semi-supervised approaches on two public pancreatic segmentation datasets, demonstrating enhanced accuracy, generalizability, and robustness.

Abstract

Accurate pancreas segmentation is critical for early cancer diagnosis, where annotation scarcity necessitates Semi-Supervised Learning (SSL). However, due to significant inter-sample morphological variability, existing SSL methods face severe generalizability limitations under sparse supervision, leading to the Supervision Bias problem. To address this, we propose Structural Consensus-based KAN Prototype Learning (SCKAN), which constructs the first cross-sample structural consensus learning with Kolmogorov-Arnold Networks (KANs), to achieve more generalizable and accurate segmentation. Specifically, SCKAN contains two key designs: Structure-constrained Prototype Consistency Learning (SPCL), which prompts unbiased structural representation by enforcing cross-sample consistency via prototype-level contrastive optimization, and Consensus-based Kolmogorov-Arnold Fusion (CKaF), which reduces morphology-specific bias by aggregating stable consensus and filtering sample-wise noise via KAN's adaptive B-spline nonlinearity. Extensive experiments on two public pancreas datasets demonstrate the effectiveness of SCKAN. Code is at https://github.com/rhodaliu17/SCKAN.

Problem

• pancreas segmentation
• semi-supervised learning
• supervision bias
• morphological variability
• generalizability

Innovation

• Kolmogorov-Arnold Networks
• semi-supervised segmentation
• structural consensus
• prototype learning
• pancreas segmentation