CAJ | 학술논문

图像匹配是计算机视觉中众多问题的一个基本方面,包括目标和场景识别,从多图片里解决三维结构,立体相关性和动作追踪。本论文主要对SIFT算法的基本步骤进行了简单阐述,并且阐述了其在场景分类上的独特应用,以及一些重要改进。实验结果表明该算法具有在场景识别和分类上有较显著的优势,可以进行良好应用。
도상필배시계산궤시각중음다문제적일개기본방면,포괄목표화장경식별,종다도편리해결삼유결구,입체상관성화동작추종。본논문주요대SIFT산법적기본보취진행료간단천술,병차천술료기재장경분류상적독특응용,이급일사중요개진。실험결과표명해산법구유재장경식별화분류상유교현저적우세,가이진행량호응용。
Image matching is a fundamental aspect of many problems in Compter Vision,including object and scene recognition,solving 3D structure in more images,3D correlation and motion tracking.The paper is mainly about SIFT alorithm and the its application in scenes classification and some of its improvement.The experiments results shows that it has significant advantages in scene recognition and classification,it also can be improved and adopted in other computer vision.