Research On Quality Evaluation And Identification Method Of Plier Cutting Marks In Forensic Science

In forensic science,the plier toolmark is common mark in crime scene.For investigator,the identification of the plier toolmark at the crime scenes can provide important clues for case-investigation and evidence for the court-decision.The morphology of the plier toolmark is complex,and the identification through manual comparison is full of subjectivity.An objective and quantitative identification method is needed.The establishment of the identification-method needs to solve two problems: one is how to determine the identificable area on the morphology of the plier toolmark;the second is how to describe the degree of identification of the plier toolmark quantitatively.By analyzing the process of forming plier toolmark,it is found that the plier toolmark is formed when the cutting-edge cuts the object,and the cutting process forms stripes similar to the texture of the cutting edge,of course,as well as the fracture and accumulation of the stripes.The stripe characteristics of the plier toolmarks can reflect texture feature of the edge.If a certain position of a plier edge cut different objects continuously,stripe texture of these objects surface may be more similarity,with concentrated reflection of plier-edge texture characteristics of the corresponding location.The stripe feature is the quasi-characteristic of this set of plier toolmark.In the appearance of plier toolmark,stripe features are the key to the recognition,so stripe features should be strengthened before the recognition of traces.Based on the need of recognition process,it is needed to preprocess the original morphology of plier toolmark,for the purpose of strengthening the fringe characteristics by establishing the preprocessing program of plier toolmark.In this paper,the theory of full element matching(CMC)is researched.Based on its stripe characteristics,the plier toolmark is divided into flat rectangular elements along the stripe direction,to separate the valid and invalid correlation regions in the topography.By looking for the consistency between the cell elements,the accuracy of the recognition of the plier toolmark is improved.As a result,the CMC method which is used for identifying the plier toolmark is established.By this method,we can distinguish 19 in 20 pairs of pliers correctly containing total 400 samples of plier toolmark,which can be regard as a great recognition effect.In this paper,a quality-evaluation method based on self-correlation is proposed.The toolmark quality is evaluated before the toolmark recognition,and the qualitycontinuity map is established.Finally,the relationship between the quality-continuity map and the recognition score is established by building the Res Net neural network,and the recognition accuracy of the neural network could reach 79.72%.The results showed that the quality-evaluation method could separate the high-quality samples from low ones with an accuracy of 79.72%.Finally,the recognition rate of trace samples increased from 75.91% to 81.01%.