Application Of The ROI (Region Of Interest) Color Method Using A Bounding Box Approach To Detect Caries On Dental Images

Abstract

Caries detection system on radiographic image detection is widely used in medicine, especially dentistry. Dentists currently perform caries detection. It becomes less accurate in determining the diagnosis because of the limitations of human vision, then needs something to help the system detect dental caries in the image. This system could be used more effectively and efficiently than with manual detection. Detection of the dots is one of the high-contrast detection caries (tooth unhealthy) and could be one substitute detection manually by a dentist. Teeth image processed with some image processing stages such as grayscale, binarization with men over image into two values​​: black and white, morphology (thresholding) bright object floating on a dark background, ROI (Region of Interest) Color, and bounding box approach (to label images which has a white contrast to facilitate detection). Image of the tooth that has been grayscale and binarization, then do the ROI (Region of Interest) with the Color Boundingbox approach. Training results using test data 10 tooth image system has an accuracy rate of 80% is good enough.