It is my pleasure to announce that our new article “Uncertainty-Aware Annotation Protocol to Evaluate Deformable Registration Algorithms” (first author: Loïc Peter) has been accepted for publication in IEEE Transactions on Medical Imaging, accepted for publication. The articles digs into how landmark correspondences, which are often used as gold standard to evaluate image registration algorithms, can bias results depending on how landmarks to annotate are selected. The article introduces a framework based on Gaussian processes that addresses this issue by: (i) iteratively suggesting the most informative location to annotate next, taking into account its redundancy with previous annotations; (ii) extending traditional pointwise annotations by accounting for the spatial uncertainty of each annotation; and (iii) providing a new strategy for the evaluation of deformable registration algorithms.
The code is publicly available at https://github.com/LoicPeter/evaluation-deformable-registration. The paper can be found here or here.
Thanks, Loïc (and Danny, and Caroline), for all the hard work!