Dental caries detection and segmentation with DELKA algorithm based on machine learning

Dental examinations and diagnosis of dental caries have been done by radiography for almost a century. However, radiography is limited in the detection of occlusal surface caries. More importantly, the use of radiographs is not required for any examination and can be very harmful to the body. For this reason, in this paper we have designed a new algorithm and system called DELKA, that segments the image into 7 areas with different levels including high decay, low decay, discolored and healthy areas by ultraviolet light irradiation and by using fluorescence images of the teeth. This segmentation is based on the k-means algorithm. By studying the output of the DELKA system, you will find that the DELKA system has been able to segment the decayed and healthy areas well.