Query by image medical training: optical biopsy with confocal endoscopy (OB-CEM)

Abstract

The use of Optical Biopsies-OB (in the present case Confocal endomicroscopy-CEM) is limited due to difficulties to interpret images. The OB-CEM are taken by endoscopists, not trained in microscopic morphology which is the domain of the surgical pathology. To gain diagnostic confidence the endoscopists could consult the images to a pathologist or could use the technique proposed in the paper. That is, to search for similar images on Internet to compare the diagnosis. The present paper is a positioning paper of how to build a CEM-image metadata to be used by the multimedia standards ISO-15938-12:2008 and ISO-24800-3 in order to search on line using a “query by image”. Metadata semantics based on Kudo colorectal crypt architecture was used for annotation or automatic image extraction. The training set was composed of 25 OB-CEM chromo-colonoscopy images taken with a FICE (Fujinon Intelligent Chromoendoscopy). Those parameters were, whenever possible, automatically extracted from the image and included in the metadata for image mining. Future developments will annotate histological images is such a way that the query could also retrieve the histological image.