Additive Causal Construction for Transferable and Reconfigurable Cross-System Learning in Multi-Source Image Fusion

2026-06-30

AI Summary

This work addresses the significant degradation in out-of-distribution generalization of multi-source image fusion models caused by cross-system discrepancies (CSD) and entanglements (CSE) arising from system heterogeneity. To mitigate this, the authors propose the Additive Causal Construction (ACC) framework, which unifies transferability and reconstructability within a causal graph for the first time. ACC establishes shared causal “anchors” across systems via intervention consistency to enable effective cross-system transfer, while modeling the fusion process as a causal construction with explicit uncertainty quantification. By integrating content-mechanism disentanglement and response alignment, the framework enhances robustness. Experiments on ColorMNIST and multi-center medical imaging tasks for MVI prediction demonstrate that ACC substantially improves out-of-distribution generalization without compromising strong in-distribution performance.

Abstract

In multi-source image fusion scenarios, heterogeneous inputs are typically driven by distinct generative mechanisms and can be viewed as a composition of multiple causal systems. However, cross-system discrepancy (CSD) and cross-system entanglement (CSE) commonly arise during the fusion process, often leading to significant performance degradation under out-of-distribution (OOD) predictions. To address the CSD and CSE issues, we propose the additive causal construction (ACC) framework, which characterizes information fusion at two levels: firstly, it establishes causal "anchors" shared among multiple systems through intervention consistency to enable causal graph transferability (CGT); and secondly, it formalizes the fusion process as causal construction and models the reliability of constructed paths through uncertainty quantification to ensure causal graph reconfigurability (CGR). Building upon this, we revisit the traditional causal representation learning (CRL) with ACC and propose ACC-CRL as a learnable instantiation of the framework. The method explores joint causal content representations across systems via content-mechanism decoupling, and performs response alignment under shared anchors to mitigate CSD. Furthermore, it incorporates structural uncertainty to adaptively regulate the fusion process, thereby suppressing unstable CSE. We conduct systematic experiments on synthetic data (ColorMNIST) and real-world multi-center medical imaging tasks (microvascular invasion (MVI) prediction). The results demonstrate that the proposed method significantly improves OOD generalization while maintaining in-distribution (ID) performance, validating the effectiveness and robustness of the ACC-CRL strategy based on mechanism alignment and uncertainty modeling in open environments.

Problem

• cross-system discrepancy
• cross-system entanglement
• out-of-distribution generalization
• multi-source image fusion
• causal representation learning

Innovation

• Additive Causal Construction
• Causal Graph Transferability
• Causal Graph Reconfigurability
• Uncertainty Quantification
• Content-Mechanism Decoupling