基于卷积神经网络的钳剪工具痕迹识别

钳剪工具痕迹识别对法庭审判和侦查破案有着重要的参考价值，是物证分析识别的重要组成部分。针对该类工具种类繁多，现场痕迹复杂多样的特点，本文提出了一种基于卷积神经网络识别的钳剪痕迹分析方法。使用断线钳、线缆钳等10类常用钳剪工具，采集制作了300枚钳剪样本，在此基础上对特征区域进行录制，共200余段视频，提取钳剪痕迹特征图像共120 000张。提出TpsNet，以钳剪断头的侧面图片为识别分类对象，通过图片的分类实现对钳剪痕迹的分析识别。结果表明，TpsNet模型在钳剪痕迹数据集上的分类精度达到97.56 %，可作为钳剪痕迹分析与识别的重要依据。

Plier and Scissor Mark Recognition Based on Convolutional Neural Network

The recognition of pliers and scissors marks has important reference value for court trials and investigations. It is an important part of the analysis of criminal evidence. Aiming at the variety of tools and the complex and diverse traces of the field, this paper proposes a method based on convolutional neural network to recognize the plier and scissor marks. 300 cutting sample were made from 10 common pliers and scissors such as bolt cutters and cable cutters. More than 200 videos in the feature area were recorded and 120000 cutting sample feature images were extracted. In this paper. TpsNet is proposed. The side picture of the pliers and scissors is used as the recognition object to realize the pliers and scissors. Marks are recognized through the classification of the pictures. The results show that the classification accuracy of the TpsNet model on the data set reaches 97.56%. It is concluded that it can be used as an important basis for the analysis and recognition of the plier and scissor marks.