Unsupervised training of keypoint-agnostic descriptors for flexible retinal image registration

2025-05-05

AI Summary

Supervised registration of color fundus images suffers from scarce annotated data and poor generalizability due to reliance on hand-crafted or learned keypoint detectors. Method: This paper proposes a keypoint-agnostic unsupervised descriptor learning framework that decouples descriptor learning from keypoint detection. It jointly optimizes feature representations via contrastive learning and reconstruction constraints, eliminating the need for any keypoint detector during both training and inference. To enhance robustness evaluation, we introduce a multi-detector consensus verification mechanism and a lightweight, custom-designed detector module. Results: On public fundus datasets, our method achieves registration accuracy comparable to supervised approaches (mean TRE < 5.2 pixels) and demonstrates consistent performance across diverse detectors—including SIFT, ORB, and SuperPoint—thereby significantly improving the generality and clinical applicability of medical image registration.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Unsupervised & Self-Supervised 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
• Security and Privacy: Large-scale security measurements
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Current color fundus image registration approaches are limited, among other things, by the lack of labeled data, which is even more significant in the medical domain, motivating the use of unsupervised learning. Therefore, in this work, we develop a novel unsupervised descriptor learning method that does not rely on keypoint detection. This enables the resulting descriptor network to be agnostic to the keypoint detector used during the registration inference. To validate this approach, we perform an extensive and comprehensive comparison on the reference public retinal image registration dataset. Additionally, we test our method with multiple keypoint detectors of varied nature, even proposing some novel ones. Our results demonstrate that the proposed approach offers accurate registration, not incurring in any performance loss versus supervised methods. Additionally, it demonstrates accurate performance regardless of the keypoint detector used. Thus, this work represents a notable step towards leveraging unsupervised learning in the medical domain.

Problem

• Develop unsupervised descriptor learning for retinal images
• Enable keypoint-agnostic registration without labeled data
• Validate performance across diverse keypoint detectors

Innovation

• Unsupervised descriptor learning without keypoint detection
• Keypoint-agnostic network for flexible registration
• Accurate performance across varied keypoint detectors