LandMatch: A collaborative annotation Tool for Landmark Matching and Benchmarking 3D volumes Registration in Longitudinal Studies

Accurate registration is an important step in longitudinal medical image studies.However, the scarcity of ground-truth annotations limits the reproducibility and benchmarking of the state-of-the-art methods.We introduce a lightweight, collaborative annotation tool for identifying the corresponding points in 3D longitudinal scans.Designed for radiologists to easily use, the tool enables a precise and scalable evaluation of registration algorithms through ground-truth landmarks.We demonstrate its application on the brain MRI longitudinal dataset BraTS-Reg and report benchmark metrics across standard registration methods.