Useful GLCM textural properties in the classification of colonic mucosa microscopic images

• Main Authors: Gan Lim, Laurence A., Naguib, Raouf N. G., Dadios, Elmer P., de la Fuente, Debbie
• Format: text
• Published: Animo Repository 2007
• Subjects: Colon (Anatomy)—Cancer—Imaging; Three-dimensional imaging; Computer Engineering
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/5832
• Institution: De La Salle University

This paper reports about extraction and analysis of textural features of colonic mucosa microscopic images. The data presented here is a preliminary result of a much larger study on automatic classification of colonic mucosa microscopic images using textural features and AI structures proposed by Gan Lim et al. (2007). The images used were initially classified by a human expert into three classifications: normal, neoplastic, and malignant. A total of 14 features were considered and analysis of the features showed that the mean, correlation, sum average, and sum variance were more effective in discriminating the images compared to other GLCM-derived properties.