Semi-Automatic Integrated Segmentation Approaches and Contour Extraction Applied to Computed Tomography Scan Images

摘要

We propose to segment two-dimensional CT scans traumaticbrain injuries with various methods. These methods arehybrid, feature extraction, level sets, region growing, andwatershed which are analysed based upon their parametricand nonparametric arguments. The pixel intensities, gradientmagnitude, affinity map, and catchment basins of thesemethods are validated based upon various constraints evaluations.In this article, we also develop a new methodology fora computational pipeline that uses bilateral filtering, diffusionproperties, watershed, and filtering with mathematicalmorphology operators for the contour extraction of the lesionin the feature available based mainly on the gradientfunction. The evaluations of the classification of these lesionsare very briefly outlined in this context and are beingundertaken by pattern recognition in another paper work.