安检CT图像目标物识别方法研究进展

Research Progress in Target Recognition Methods for Security Computed Tomography Images

基于先进的安检CT扫描设备,研究三维CT图像中的目标物识别方法,对于提升安检质量的可靠性和稳定性具有十分重要的意义。本文系统评述安检CT图像目标物识别的最新研究进展和发展趋势,不仅对近年来本领域的代表性学术研究成果进行总结回顾,还首次将工业界的相关专利也纳入分析范围;在文献归类环节,除沿用常见的技术路线和任务类型两种归类法,还根据计算空间维度的差异提出一种新的归类法,并对检索到的代表性文献进行多标签归类;通过总结分析本领域研究成果,发现3个重要特点,分析其主要成因,最后对本领域的研究方向提出几点建议。

Researching target recognition methods for 3D computed tomography(CT)images is important for improving the reliability and stability of the security inspection quality in advanced security CT scanning equipment.This article systematically reviews the latest research progress and developmental trends in target recognition of CT images for security purposes.Not only representative academic research achievements of recent years in this field are summarized and reviewed,but also relevant industrial patents are included for the first time.For literature classification,in addition to two common taxonomies,a new taxonomy is proposed,and the retrieved representative literature is classified with multiple labels.In summarizing and analyzing the research achievements in this field,three important characteristics were identified,and the origins were analyzed.Finally,some suggestions are put forward for future research directions in this field.