CAJ | 학술논문

遮挡问题在现实世界中经常发生，如何在遮挡情况下准确检测和识别感兴趣的目标是计算机视觉领域的一个重要课题。由于遮挡目标的信息不全，所以遮挡目标的检测和识别是一个具体挑战的课题，其研究具有重要意义。本文对基于图像的遮挡目标检测与识别方法按照三大类进行分类总结分析，在分析各类方法基本思路和优缺点基础上，给出了遮挡目标检测识别方法的发展趋势。
차당문제재현실세계중경상발생，여하재차당정황하준학검측화식별감흥취적목표시계산궤시각영역적일개중요과제。유우차당목표적신식불전，소이차당목표적검측화식별시일개구체도전적과제，기연구구유중요의의。본문대기우도상적차당목표검측여식별방법안조삼대류진행분류총결분석，재분석각류방법기본사로화우결점기출상，급출료차당목표검측식별방법적발전추세。
Object occlusion often appears in real world, and it is an important problem how to detect and recognize interesting objects under occlusion in computer vision research field. Due to some information loss in occluded object, it’s a chal enging task to detect and recognize occluded object. First, this paper classifies the occluded object detection and recognition methods into three categories. Then, it summarizes and analyses the fundamentals, merits and drawbacks of the methods in details. Final y, it gives out the outlook of these methods.