DeepATLAS: One-Shot Localization for Biomedical Data

arXiv:2402.09587v1 Announce Type: new
Abstract: This paper introduces the DeepATLAS foundational model for localization tasks in the domain of high-dimensional biomedical data. Upon convergence of the proposed self-supervised objective, a pretrained model maps an input to an anatomically-consistent embedding from which any point or set of points (e.g., boxes or segmentations) may be identified in a one-shot or few-shot approach. As a representative benchmark, a DeepATLAS model pretrained on a comprehensive cohort of 51,000+ unlabeled 3D computed tomography exams yields …

abstract arxiv biomedical consistent convergence cs.cv data domain embedding foundational model localization maps paper set tasks type