Telediagnostic: Specimen Matching and Retrievalfor Medical Diagnosis

Abstract

Effective management of medical disaster outbreak requires reliable diagnosis of ailing people and a trained physician to interpret the data in order to give the correct guided treatment. In many developing countries such as Kenya, bringing these two components together - reliable data and trained physicians has been a tremendous challenge. The main objective of this project was to develop a prototype of an image matching application for efficient content-based retrieval of visual data from digital sources. Using Oracle 109 database technology, a pattern comparison of the raw gathered images was to be stored and compared with a bench marked known image using a threshold based Euclidean distance (attributes, color, texture, and shape). This application demonstrated the possibility of computerizing the diagnosis of major diseases that may require discerning human eye, microscopic observation or through digital medium. Using a camera of a cell phone turned into a clinically quality light microscope or USB microscope, the captured images of patient blood samples can be download into a computer for evaluation through this intelligent content-based pattern matching application. The results are then given with minimum pathologist intervention. The cell-phone microscopy device enables visualization and capture of specimens for diagnosis.