One-Shot Medical Landmark Localization by Edge-Guided Transform and Noisy Landmark Refinement

As an important upstream task for many medical applications, supervised landmark localization still requires non-negligible annotation costs to achieve desirable performance. Besides, due cumbersome collection procedures, the limited size of datasets impacts effectiveness large-scale self-supervised pre-training methods. To address these challenges, we propose a two-stage framework one-shot localization, which first infers landmarks by unsupervised registration from labeled exemplar unlabeled targets, and then utilizes noisy pseudo labels train robust detectors. handle significant structure variations, learn end-to-end cascade global alignment local deformations, under guidance novel loss functions incorporate edge information. In stage II, explore self-consistency selecting reliable cross-consistency semi-supervised learning. Our method achieves state-of-the-art performances on public different body parts, demonstrates its general applicability. Code is available at https://github.com/GoldExcalibur/EdgeTrans4Mark .