Image Processing For Body-surface Trauma In Forensic Authentication

As for the authentication of body-surface trauma, no uniform and standardized methods for the measurement have been found yet, since in most cases the body surfaces and trauma shapes are anomalistic. Traditional methods have trivial and complicated manipulation and non-accurate results, which are unable to meet the requirements of judicial authentication. It also apt to be interfered by subjective or man-made factors, which often leads to iterative authentications by different forensic doctors and quarrels on the inconsistent results, and to some degree influences accurate implement of relevant criterion and law. In addition, these methods are mostly contact-manipulation. As for fresh wounds, these methods are not conformable to the clinic diagnosis and treatment, and also inconvenient for mensuration of trauma in certain positions. Forensic authentication using image-processing methods is advantageous to get better measurements and authentication of the complex body surface trauma and non-normally trauma shapes, which result in various trauma images. In this dissertation, a new image-processing platform to analyze the trauma images has been set up, and the image-processing methods for trauma images have been discussed detailedly.The trauma image-processing platform is developed using Visual C++ 6.0(Microsoft) programming language. It is an integrated software package, which integrates image acquiring, files managing, image processing, parameter measuring, result outputting and report processing together. Because OrientalObject Programming (OOP) technology is applied to this system, the modules in this package are easy to accomplish functional expanding and code implanting.Image segmentation plays a key role in image analysis and it is also the most difficult step in image processing procedure. As for those images that have acute contrast in colors between the trauma part and the background, a method that has four steps based on the segmentation skills for color images is proposed. Image enhancing, color adjusting, RGB heft extracting, mathematical morphology processing are implemented according to the set sequence. As for gray-average images, another method that has seven steps based on Snakes model is put forward. These steps in turn are image enhancing, image graying, median filter, edge smoothing, initial contour extraction, snakes model iteration with different parameters, and B sample imitation. As for seriously gray changing images, another method which has four steps based on thresholding is brought forward. Image enhancing, RGB heft or combination extracting,median filter, adaptable threshold segmentation according to image feature.Due to the wide range of body-surface trauma parametersx, seven kinds of trauma parameters, curve area measuring, closing polygon area measuring, pixel staticstics area measuring, code area measuring, curve cutting length measuring, polygon cutting length measuring, and angle measuring can be got in our program.An authentication report plays a serious role in judicial appraisal. Six kind of authentication reports are designed according to the judicial requirement, which are respectively related to the authentication of damnify, pathology, toxicology, psychiatry, clinic, and material evidence.Truant images are processed through this image-processing platform. Trauma images are segamented by distinguish various adoptable techniques, half-auto or changing between users and machines extracting contours which inosculate actual boundarys more. The functions of computing trauma parameters, accumulating total trauma parameters, filling in a great variety of authentication reports are fully implemented. The results indicate that the image processingsystem can be useful for the quantitative analysis of trauma parameters with rapidity and accuracy.