New Method To Improve The Quality Of Data Recording Odontogram Through Occlusal Dental Photography For Forensic Odontology
The problem in this research is integration the conventional odontogram data recording method into a computer-based odontogram recording through occlusal dental photography in accordance with the National Standard of Dental Medical Records. The purpose of this study is to make the dental examination process more accurate, to evaluate the patient's dental biometrics with occlusal photo data of the patient, and to provide odontogram data as a quality and accurate antemortem data for identification purposes, so that data is evaluated and used as material for discussion the doctors as well as for the patient's family if needed. The method proposed in this study is a simulation using MATLAB by testing binary images then segmenting with Watershed transformation and morphological processes to obtain a single tooth image so that the process is obtained from the numbering process according to FDI (International Dental Federation) standards, then testing missing teeth detection, accuracy of teeth position and calculate classification accuracy using the K-NN (K-Nearest Neighbor) method. The results of this study were obtained odontogram form for maxilla fit accuracy at 73.6%, while mandible accuracy at 100%, the collaboration results of Color Moment and HOG feature extraction in this new method system obtained the best performance with using positive data, occurs at K = 1 of 98%, with an average computing time of 43.95 seconds. While classification using negative data, the best performance is 50%, occurring at K = 3 and K = 7, with an average computing time of 4.35 seconds.