Comparison of Stereo Matching Algorithms for the Development of Disparity Map

Q4 Engineering

International Journal of Computer Aided Engineering and Technology Pub Date : 2023-03-31 DOI:10.24297/ijct.v23i.9390

Hamid Fsian, Vahid Mohammadi, Pierre Gouton, Saeid Minaei

{"title":"Comparison of Stereo Matching Algorithms for the Development of Disparity Map","authors":"Hamid Fsian, Vahid Mohammadi, Pierre Gouton, Saeid Minaei","doi":"10.24297/ijct.v23i.9390","DOIUrl":null,"url":null,"abstract":"Stereo Matching is one of the classical problems in computer vision for the extraction of 3D information but still controversial for accuracy and processing costs. The use of matching techniques and cost functions is crucial in the development of the disparity map. This paper presents a comparative study of six different stereo matching algorithms including Block Matching (BM), Block Matching with Dynamic Programming (BMDP), Belief Propagation (BP), Gradient Feature Matching (GF), Histogram of Oriented Gradient (HOG) and Fixed-Window Aggregated Cost (FWAC). In addition, three cost functions, namely, Mean Squared Error (MSE), Sum of Absolute Differences (SAD), and Normalized Cross-Correlation (NCC) were utilized and compared. The stereo images used in this study were obtained from the Middlebury Stereo Datasets provided with perfect and imperfect calibrations. It was observed that the selection of matching function is quite important and also depends on the image properties. Results showed that the BP algorithm in most cases provided better results achieving accuracies over 95%. Accordingly, BP algorithm is highly recommended based on rapidity and performance and for applications with the need of detection of small details, HOG is advised.","PeriodicalId":38492,"journal":{"name":"International Journal of Computer Aided Engineering and Technology","volume":"283 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Computer Aided Engineering and Technology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.24297/ijct.v23i.9390","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"Engineering","Score":null,"Total":0}

引用次数: 1

Abstract

Stereo Matching is one of the classical problems in computer vision for the extraction of 3D information but still controversial for accuracy and processing costs. The use of matching techniques and cost functions is crucial in the development of the disparity map. This paper presents a comparative study of six different stereo matching algorithms including Block Matching (BM), Block Matching with Dynamic Programming (BMDP), Belief Propagation (BP), Gradient Feature Matching (GF), Histogram of Oriented Gradient (HOG) and Fixed-Window Aggregated Cost (FWAC). In addition, three cost functions, namely, Mean Squared Error (MSE), Sum of Absolute Differences (SAD), and Normalized Cross-Correlation (NCC) were utilized and compared. The stereo images used in this study were obtained from the Middlebury Stereo Datasets provided with perfect and imperfect calibrations. It was observed that the selection of matching function is quite important and also depends on the image properties. Results showed that the BP algorithm in most cases provided better results achieving accuracies over 95%. Accordingly, BP algorithm is highly recommended based on rapidity and performance and for applications with the need of detection of small details, HOG is advised.

查看原文本刊更多论文

视差图开发的立体匹配算法比较

立体匹配是计算机视觉中三维信息提取的经典问题之一，但在精度和处理成本等方面一直存在争议。匹配技术和成本函数的使用在视差图的开发中是至关重要的。本文对分块匹配(BM)、动态规划分块匹配(BMDP)、信念传播(BP)、梯度特征匹配(GF)、定向梯度直方图(HOG)和固定窗口聚合代价(FWAC)等六种立体匹配算法进行了比较研究。此外，利用均方误差(Mean Squared Error, MSE)、绝对差和(Sum of Absolute difference, SAD)和归一化相互关系(Normalized Cross-Correlation, NCC)三个成本函数进行了比较。本研究中使用的立体图像来自Middlebury立体数据集，提供了完美和不完美的校准。结果表明，匹配函数的选择非常重要，并且取决于图像的属性。结果表明，BP算法在大多数情况下提供了更好的结果，准确率超过95%。因此，基于速度和性能，强烈推荐BP算法，对于需要检测小细节的应用，建议使用HOG算法。

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

International Journal of Computer Aided Engineering and Technology Engineering-Engineering (all)

CiteScore

1.10

自引率

0.00%

发文量

期刊介绍： IJCAET is a journal of new knowledge, reporting research and applications which highlight the opportunities and limitations of computer aided engineering and technology in today''s lifecycle-oriented, knowledge-based era of production. Contributions that deal with both academic research and industrial practices are included. IJCAET is designed to be a multi-disciplinary, fully refereed and international journal.