Surrogate Supervision for Robust and Generalizable Deformable Image Registration

2025-09-11

AI Summary

Deep learning-based image registration methods exhibit limited robustness to input variations such as image artifacts, field-of-view mismatches, and cross-modality discrepancies. To address this, we propose a proxy-supervised training paradigm that decouples the input domain from the supervision domain: a network-estimated deformation field is applied to a proxy image to synthesize a deformed counterpart, which is then optimized end-to-end via a normalized similarity metric. This framework eliminates the need for ground-truth deformation fields, significantly improving training stability for heterogeneous medical images—including brain MRI, lung CT, and multi-contrast MR. Experiments demonstrate that our method maintains high registration accuracy under complex degradations while preserving performance on standard benchmarks. It thus achieves both strong generalizability and practical applicability.

Technology Category

• Computer Vision: Medical and Biological Imaging
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Objective: Deep learning-based deformable image registration has achieved strong accuracy, but remains sensitive to variations in input image characteristics such as artifacts, field-of-view mismatch, or modality difference. We aim to develop a general training paradigm that improves the robustness and generalizability of registration networks. Methods: We introduce surrogate supervision, which decouples the input domain from the supervision domain by applying estimated spatial transformations to surrogate images. This allows training on heterogeneous inputs while ensuring supervision is computed in domains where similarity is well defined. We evaluate the framework through three representative applications: artifact-robust brain MR registration, mask-agnostic lung CT registration, and multi-modal MR registration. Results: Across tasks, surrogate supervision demonstrated strong resilience to input variations including inhomogeneity field, inconsistent field-of-view, and modality differences, while maintaining high performance on well-curated data. Conclusions: Surrogate supervision provides a principled framework for training robust and generalizable deep learning-based registration models without increasing complexity. Significance: Surrogate supervision offers a practical pathway to more robust and generalizable medical image registration, enabling broader applicability in diverse biomedical imaging scenarios.

Problem

• Improving robustness of deformable image registration networks
• Enhancing generalizability across varying input characteristics
• Addressing artifacts, field-of-view and modality differences

Innovation

• Surrogate supervision decouples input and supervision domains
• Applies transformations to surrogate images for training
• Ensures similarity computed in well-defined domains