Automated Reconstruction of Dendritic and Axonal Trees by Global Optimization with Geometric Priors

摘要: We present a novel probabilistic approach to fully automated delineation of tree structures in noisy 2D images and 3D image stacks. Unlike earlier methods that rely mostly on local evidence, ours builds set candidate trees over many different subsets points likely belong the optimal then chooses best one according global objective function combines evidence with geometric priors. Since does not necessarily span all points, algorithm is able eliminate false detections while retaining correct topology. Manually annotated brightfield micrographs, retinal scans DIADEM challenge datasets are used evaluate performance our method. metric quantitatively topological accuracy reconstructions showed use regularization yields substantial improvement.