Implementation of GA-KSOM and ANFIS in the classification of colonic histopathological images

• Main Authors: Gan Lim, Laurence A., Maguib, Raouf N. G., Dadios, Elmer P., Avila, Jose Maria C.
• Format: text
• Published: Animo Repository 2012
• Subjects: Colon (Anatomy)—Cancer—Diagnosis; Self-organizing maps; Image analysis; Genetic algorithms; Mechanical Engineering
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/1721; https://animorepository.dlsu.edu.ph/context/faculty_research/article/2720/type/native/viewcontent
• Institution: De La Salle University

The WHO reports that colon cancer is one of the leading causes of cancer mortality in the world with the majority of people with this type of cancer belonging to those who are 60 years or older. Similar to other types of cancer, early detection is very important for a successful treatment. This paper reports on the implementation of Kohonen Self-Organizing Map (KSOM) with genetic algorithms (GA), and neuro-fuzzy classifier to classify colonic histopathological images into normal, adenomatous polyp, and cancerous. KSOM with GA, or GA-KSOM for short, was used in the feature selection stage while a neuro-fuzzy algorithm was used in the classification stage. ANFIS or Adaptive Neuro-Fuzzy Inference System was chosen as the structure/architecture of the neuro-fuzzy algorithm. The classification accuracies obtained were very promising with 86.7% and 87.8% for the training and testing sets, respectively. © 2012 IEEE.