Kristopher David Harjono, Gede Putra Kusuma Negara
{"title":"Object instance recognition using best increasing subsequence","authors":"Kristopher David Harjono, Gede Putra Kusuma Negara","doi":"10.1109/KICSS.2016.7951432","DOIUrl":null,"url":null,"abstract":"Object instance recognition enables the realization of many potential applications, such as information retrieval, scene understanding and human computer interaction. However, it is still a challenging problem in computer vision. The appearance of an object is affected by variations in illumination, viewpoint and occlusion. In this contribution, we propose an object instance recognition method based on Best Increasing Subsequence. It estimates a set of geometrically consistent feature pairs and at the same time, maximizes the total similarity score between test and train images. Our experimental results show that the proposed method outperforms the existing geometric verification methods, RANSAC Homography and Weighted LIS.","PeriodicalId":170692,"journal":{"name":"International Conference on Knowledge, Information, and Creativity Support Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2016-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Conference on Knowledge, Information, and Creativity Support Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/KICSS.2016.7951432","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
Object instance recognition enables the realization of many potential applications, such as information retrieval, scene understanding and human computer interaction. However, it is still a challenging problem in computer vision. The appearance of an object is affected by variations in illumination, viewpoint and occlusion. In this contribution, we propose an object instance recognition method based on Best Increasing Subsequence. It estimates a set of geometrically consistent feature pairs and at the same time, maximizes the total similarity score between test and train images. Our experimental results show that the proposed method outperforms the existing geometric verification methods, RANSAC Homography and Weighted LIS.