Title :
Automatic classification of neoplastic lesion on gastric biopsy images
Author :
Emi Morotomi;Toshiyuki Tanaka
Author_Institution :
Department of Science and Technology, Keio University, Kanagawa, Japan
fDate :
7/1/2015 12:00:00 AM
Abstract :
In histopathological diagnosis, pathologists observe the biopsy images and diagnose the tumor grade. However, the number of pathologists has been decreasing, so the demand for cancer diagnosis support system has been increasing in recent years. Therefore, this study proposes the method for automatic classification to two classes which are neoplastic lesion, and non-neoplastic lesion. Our method consists of image inputting, region extraction, feature calculation, and discriminant analysis. As the result, our method showed 93.33% accuracy on the neoplastic lesion, and 82.86% accuracy on the non-neoplastic lesion.
Keywords :
"Glands","Feature extraction","Biopsy","Lesions","Accuracy","Shape","Cancer"
Conference_Titel :
Society of Instrument and Control Engineers of Japan (SICE), 2015 54th Annual Conference of the
DOI :
10.1109/SICE.2015.7285506