Towards building computerized image analysis framework for nucleus discrimination in microscopy images of diffuse glioma

As an effort to build an automated and objective system for pathologic image analysis, we present, in this paper, a computerized image processing method for identifying nuclei, a basic biological unit of diagnostic utility, in microscopy images of glioma tissue samples. The complete analysis includes multiple processing steps, involving mode detection with color and spatial information for pixel clustering, background normalization leveraging morphological operations, boundary refinement with deformable models, and clumped nuclei separation using watershed. In aggregate, our validation dataset includes 220 nuclei from 11 distinct tissue regions selected at random by an experienced neuropathologist. Computerized nuclei detection results are in good concordance with human markups by both visual appraisement and quantitative measures. We compare the performance of the proposed analysis algorithm with that of CellProfiler, a classical analysis software for cell image process, and present the superiority of our method to CellProfiler.