Malignant colon cancer detection through image processing and application of artificial neural network

• Main Authors: Albis, Christine Joy R., Bernardo, Victor Vincent R., Wong, Marie Antonette S.
• Format: text
• Language: English
• Published: Animo Repository 2007
• Subjects: Artificial intelligence > Medical applications; Neural networks (Computer science); Colon (Anatomy) > Cancer > Diagnosis; Cancer > Computer simulation
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/5949
• Institution: De La Salle University

Cancer is one of the dreadfull diseases that persistently challenge biomedical engineering to use electronic means of detecting at an early stage of this disease. The inconsistent diagnostic results conducted in medical laboratories contribute to the demand of automatic verification system using digital image processing and artificial intelligence. The advancement of technology dictates the capability of biomedical applications to create an automatic detection system that adopts the capability of the pathologist in biopsy. This research achieved a means of detecting tissue sample images of candidate for Colon Cancer as positive or negative of malignancy. This research uses digital image processing and artificial neural network to support the qualitative assessment of the pathologists. It uses the different microscopic characteristics of colon cancer images such as Nuclei Formation, Presence of Lumen, Nuclei vs. Cytoplasm Ratio and Uniformity, in classifying colon cancer images as positive or negative of malignancy. Applying the concept in the techniques used in this study, the system was able to correctly detect several tissue sample images as positive or negative of Colon Cancer with an accuracy of 92.4%.