Object segmentation and shape reconstruction using computer-assisted segmentation tools

Quantitative analysis of digital images requires detection and segmentation of the borders of the object of interest. Accurate segmentation is required for volume determination, 3D rendering, radiation therapy, and surgery planning. Once an accurate segmentation is obtained, this information may be used by the radiologist to compare the volume and morphology characteristics of each region against known anatomical norms, other regions in the same image set, and corresponding regions in related image sets. In medical images, segmentation has traditionally been done by human experts. Even with the aid of image processing software (computer-assisted segmentation tools), manual segmentation of 3D CT images is tedious, time-consuming, and thus impractical, especially in cases where a large number of objects must be specified. Many methods have been proposed to detect and segment 2D shapes, the most of which is template matching. However their low speed has prevented its wide spread use. Other techniques called snakes or active contours have been used, but the main drawbacks associated with their initialization and poor convergence to boundary concavities limit their utility. Substantial computational and storage requirements become especially acute when object orientation and scale have to be considered. Therefore automated or semi-automated segmentation techniques are essential if these software applications are ever to gain widespread clinical use. In this work we will present an effective semi-automatic method, based on the boundary tracking technique, which improves the time when one or more structures are in use. Different segmentation techniques would be proposed for the particular organs of interests (lungs, skin and spine canal) and a 3D shape reconstruction of these regions would be illustrate the efficiency of the segmentation techniques. Finally, the proposed technique would be compared with the manual segmentation obtained from the doctor experts using quantitative (shape matching measures) and qualitative (visual comparison) measures.