Structural Morphology Based Automatic Virus Particle Detection Using Robust Segmentation and Decision Tree Classification | Abstract

ISSN ONLINE(2319-8753)PRINT(2347-6710)

Special Issue Article Open Access

Structural Morphology Based Automatic Virus Particle Detection Using Robust Segmentation and Decision Tree Classification

Abstract

Accurate and automatic approach to locate virus particles in electron microscopy is cardinal because of the large number of electron views that are needed to perform high resolution three dimensional reconstructions at the ultrastructural level. This paper describes a fully automatic approach to locate adenovirus particles where low level of entropy is compared to the surrounding unorganized area. Characterization of the structural morphology of the virus particles based on area and eccentricity helps to detect the candidate points. The detected points are subjected to credibility test based on features extracted from each point from a texture image followed by decision tree classification. Final validation of approved candidate’s takes place with 3D entropy proportion coordinates, computed in the original image, compensated work image1 and strongly filtered work image 2.

Saffna Shajahan, Chithra B

To read the full article Download Full Article