Local Feature Detectors Performance Analysis on Digital Image

2019 IEEE International Scientific-Practical Conference Problems of Infocommunications, Science and Technology (PIC S&T) Pub Date : 2019-10-01 DOI:10.1109/PICST47496.2019.9061331

Kirill Smelyakov, Dariia Tovchyrechko, Igor Ruban, A. Chupryna, O. Ponomarenko

{"title":"Local Feature Detectors Performance Analysis on Digital Image","authors":"Kirill Smelyakov, Dariia Tovchyrechko, Igor Ruban, A. Chupryna, O. Ponomarenko","doi":"10.1109/PICST47496.2019.9061331","DOIUrl":null,"url":null,"abstract":"The article puts an experiment on application of widely used ORB, SIFT and SURF feature point detectors represented by the corresponding functions in the OpenCV library, to the images of the most common object classes such as human faces, fine details and artificial images. The considered detectors display as a result a huge number of points that are not classified or structured. Building an appropriate classifier would greatly increase the efficiency of subsequent image processing operations: localization, recognition, search, and tracking of objects. The article analyzes the effectiveness of the experimental results at a quantitative and qualitative level taking into account the conditions and limitations (primarily temporal) on solving practical problems in the era of Big Data, as well as taking into account the fact that some detectors are proprietary. According to the analysis results of the usage effectiveness of the features points detectors considered in the work the practical recommendations for specific use cases are given at the end of the work.","PeriodicalId":6764,"journal":{"name":"2019 IEEE International Scientific-Practical Conference Problems of Infocommunications, Science and Technology (PIC S&T)","volume":"6 1","pages":"644-648"},"PeriodicalIF":0.0000,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"11","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 IEEE International Scientific-Practical Conference Problems of Infocommunications, Science and Technology (PIC S&T)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/PICST47496.2019.9061331","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 11

Abstract

The article puts an experiment on application of widely used ORB, SIFT and SURF feature point detectors represented by the corresponding functions in the OpenCV library, to the images of the most common object classes such as human faces, fine details and artificial images. The considered detectors display as a result a huge number of points that are not classified or structured. Building an appropriate classifier would greatly increase the efficiency of subsequent image processing operations: localization, recognition, search, and tracking of objects. The article analyzes the effectiveness of the experimental results at a quantitative and qualitative level taking into account the conditions and limitations (primarily temporal) on solving practical problems in the era of Big Data, as well as taking into account the fact that some detectors are proprietary. According to the analysis results of the usage effectiveness of the features points detectors considered in the work the practical recommendations for specific use cases are given at the end of the work.

查看原文本刊更多论文

数字图像局部特征检测器性能分析

本文以OpenCV库中相应的函数为代表，对广泛使用的ORB、SIFT和SURF特征点检测器在人脸、精细细节和人工图像等最常见的对象类图像中的应用进行了实验。所考虑的检测器结果显示了大量未分类或未结构化的点。构建一个合适的分类器将大大提高后续图像处理操作的效率:对象的定位、识别、搜索和跟踪。考虑到大数据时代解决实际问题的条件和限制(主要是时间上的)，以及一些检测器是专有的，本文从定量和定性两个层面分析了实验结果的有效性。根据工作中所考虑的特征点检测器的使用有效性分析结果，在工作的最后给出了针对具体用例的实用建议。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

2019 IEEE International Scientific-Practical Conference Problems of Infocommunications, Science and Technology (PIC S&T)

自引率

0.00%

发文量