GoodPoint: unsupervised learning of keypoint detection and description
This addresses the problem of keypoint detection in low-feature images like medical retina scans, though it is incremental as it builds on existing architectures like SuperPoint.
The paper tackles unsupervised learning of keypoint detection and description by training a model on homographically transformed images, achieving similar performance to SuperPoint on natural images and better performance on retina images with low corner-like features.
This paper introduces a new algorithm for unsupervised learning of keypoint detectors and descriptors, which demonstrates fast convergence and good performance across different datasets. The training procedure uses homographic transformation of images. The proposed model learns to detect points and generate descriptors on pairs of transformed images, which are easy for it to distinguish and repeatedly detect. The trained model follows SuperPoint architecture for ease of comparison, and demonstrates similar performance on natural images from HPatches dataset, and better performance on retina images from Fundus Image Registration Dataset, which contain low number of corner-like features. For HPatches and other datasets, coverage was also computed to provide better estimation of model quality.