Accuracy of Bite Mark Overlays Made by Photo Superimposition inForensics Using a Reconstructed Model of Human Being

摘要

This research aims at finding a more efficient method using computerized image superimposition with MATLAB Software in identification of bite-marks on a reconstructed model of human being. Total 34 participants were used as the sample of the study. Their bite-marks on the wax were photographed and the impressions of their upper jaws were made. Then an algorithm was designed obtaining 33 characteristics from each sample using MATLAB neural network tool box. To match the waxes and the impressions MATLAB artificial neural network tool box were used and the results were reported accordingly. Of the 34 sample impressions fed in the artificial neural network, the system could match 94% (32) cases properly with the bitten waxes. Based on the obtained results, the efficiency of photo superimposition made with MATLAB was shown.