인공신경망 기법을 이용한 상흔 분석 및 위협체 형상 예측

초록

Wound evidence after injury produced by the sharp-shaped threat pierces on a human body is closely related to the direct threat elements such as threat shape, impact energy, and incident angle, and therefore the technology to predict an initial threat shape based on the causality is very essential to solving a criminal event. In this study, the damage patterns according to various elements of the threat were analyzed, and a forensic investigation method that can predict the threat shape caused by the wound evidence was proposed. For this purpose, threat conditions including the threat shape, the impact energy, and the incident angle of the threat were set, and the changes of the area of wound evidence and the ratio of long axis length to short axis length produced in ballistic gelatin arranged for the object were constructed in the database. In addition, some arbitrary damage information was entered into the database, and the threat shape caused by the damage was investigated by means of the Fuzzy C- means clustering technique based on the probabilistic clustering algorithm. Moreover, the correlation coefficient between the input damage information and sectional shape of the detected threats was derived, and the correspondence was identified through the cross correlation analysis technique as a shape similarity evaluation method. The validity of the prediction method was verified through the example such as the two types of threat shape were specified by entering the damage information obtained through the arbitrary setting of the threat conditions, and then it was confirmed that the correlation coefficient of the threat that caused the damage (0.94) was derived higher than the other threat (0.71) by comparing the correlation coefficients. The prediction method of the threat shape proposed in this study may be used as the basis of physical prediction technique in a future forensic investigation.

키워드