DEVELOPMENT OF HISTOPATHOLOGICAL INFORMATION DATABASE SYSTEM WHICH ENABLES IMAGE RETRIEVAL USING IMAGE FEATURES AND TEXT INFORMATION

Abstract

Pathological diagnostic reports are created after the result of diagnosing specimen images which may contain cancerous tissues in histopathological diagnosis. We have developed a histopathological information database system which stores and retrieves both diagnostic reports as text information and histological samples as images. Similar images are retrieved in the database using feature vectors made by Wavelet transformation and neural network. In addition, diagnostic probability based on Bayes' theorem was created after digitization of diagnostic data with text mining technique. Then the accuracy of image retrieval is improved by the diagnostic probability model linked with images.