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

Sep 21, 2026

Uploaded by Violet

Multi-source image fusion often combines heterogeneous inputs generated by distinct causal mechanisms, where cross-system discrepancy (CSD) and cross-system entanglement (CSE) emerge during fusion and cause large performance drops under out-of-distribution (OOD) predictions. An additive causal construction (ACC) framework is proposed to address both issues by building shared causal anchors via intervention consistency for causal graph transferability, then modeling fusion as causal construction with uncertainty quantification for causal graph reconfigurability. ACC-CRL implements learnable mechanism-aligned representations using content–mechanism decoupling and structural uncertainty. Experiments on ColorMNIST and real multi-center microvascular invasion (MVI) prediction show improved OOD generalization while preserving in-distribution performance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Abstract

Index Terms

I. Introduction

Additive Causal Construction for Transferable and Reconfigurable Cross-System Learning in Multi-Source Image Fusion Zhizhong Fu 1 ,∗ , Wei Zhou 1 ,5 ,7 ,∗ , Zhaoyang Jiang2 , Yudong Lin 1 , Yifu Hou3 , Yuke Cao4 , Xiaorong Ding 1 , Qiang Yan6 , and Yifan Chen 1 arXiv :2607 .02572v1 [ cs .CV] 30 Jun 2026 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 crosssystem 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. Index Terms—Multi-source fusion, multiple causal systems, additive causal construction (ACC), cross-system discrepancy (CSD), cross-system entanglement (CSE), causal graph transferability (CGT), causal graph reconfigurability (CGR), out-ofdistribution (OOD) generalization, microvascular invasion (MVI) ∗Zhizhong Fu and Wei Zhou contributed equally to this work. Corresponding author: Yifan Chen ([yifan.chen@uestc.edu.cn](yifan.chen@uestc.edu.cn)). 1 School of Life Science and Technology, University of Electronic Science and Technology of China, Chengdu 610054, China. 2 School of Health and Wellbeing, University of Glasgow, Glasgow G12 8TB, United Kingdom. 3Department of Organ Transplantation, Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, Chengdu, China. 4 School of Information and Software Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China. 5Department of Radiology, Huzhou Maternity & Child Health Care Hospital, Huzhou 313000, China. 6Hepatological Surgery Department, Huzhou Central Hospital, Fifth School of Clinical Medicine of Zhejiang Chinese Medical University, Huzhou 313000, China. 7Department of Radiology, Huzhou Maternity & Child Health Care Hospital Affiliated to Huzhou Normal University,Huzhou 313000,China. prediction I. INTRODUCTION Multi-source data are often generated under distinct principles and can be formalized as multiple causal systems with heterogeneous mechanisms. In this context, the central challenge of image fusion extends beyond simple information aggregation to addressing heterogeneity in mechanisms. Firstly, the absence of a unified reference for causal mechanisms across systems makes semantic alignment difficult, leading to the cross-system discrepancy (CSD) issue. Secondly, the fusion process may intro

FAQs

What are the main problems addressed in multi-source image fusion?

The work focuses on cross-system discrepancy (CSD) and cross-system entanglement (CSE), which arise from heterogeneous generative mechanisms and can degrade performance under out-of-distribution (OOD) shifts.

How does the proposed ACC framework improve robustness across systems?

ACC characterizes fusion in two steps: it establishes shared causal anchors through intervention consistency to enable causal graph transferability, and it models fusion with uncertainty quantification to ensure causal graph reconfigurability.

What datasets and tasks are used to validate the approach?

Experiments are conducted on synthetic data using ColorMNIST and on real-world multi-center medical imaging for microvascular invasion (MVI) prediction.

Sep 21, 2026

Uploaded by Violet