Hierarchical, model‐based merging of multiple fragments for improved three‐dimensional segmentation of nuclei

Abstract

Background: Automated segmentation of fluorescently labeled cell nuclei in three-dimensional confocal images is essential for numerous studies, e.g., spatiotemporal fluorescence in situ hybridization quantification of immediate early gene transcription. High accuracy and automation levels are required in high-throughput and large-scale studies. Common sources of segmentation error include tight clustering and fragmentation of nuclei. Previous regionbased methods are limited because they perform merging of two nuclear fragments at a time. To achieve higher accuracy without sacrificing scale, more sophisticated yet computationally efficient algorithms are needed. Methods: A recursive tree-based algorithm that can consider multiple object fragments simultaneously is described. Starting with oversegmented data, it searches efficiently for the optimal merging pattern guided by a quantitative scoring criterion based on object modeling. Computation is bounded by limiting the depth of the merging tree.

A hyperfused mitochondrial state achieved at G 1 –S regulates cyclin E buildup and entry into S phase

“…Within the merged 3D stack, mitochondrial elements were segmented (31). After segmentation, surface points were extracted for each identified mitochondrial object in the 3D image.…”

Associative image analysis: A method for automated quantification of 3D multi-parameter images of brain tissue

“…The 3D watershed algorithm was used on the gradient-weighted distance map to break up the CyQuant channel image into nuclear fragments. A model based merging algorithm assembled these fragments into image objects representing nuclei (Lin et al, 2005b). For each object, a confidence metric was computed reflecting the quality of segmentation, represented by the goodness-of-fit of each object to pre-computed object models.…”

High content image cytometry in the context of subnuclear organization

“…Therefore, instead of rejecting these incorrectly segmented features from the analysis their recognition could also be used for enhanced segmentation. Several groups have proposed methods to merge (39,40) or separate (41,42) incorrectly segmented nuclei or groups of nuclei and by doing so, increase yield, while maintaining relatively high accuracy. The modularity and open character of the presented analysis allows future incorporation of such improvements.…”