基于信念传播的运动估计

14th International Conference on Image Analysis and Processing (ICIAP 2007) Pub Date : 2007-09-10 DOI:10.1109/ICIAP.2007.88

Giuseppe Boccignone, A. Marcelli, Paolo Napoletano, M. Ferraro

{"title":"基于信念传播的运动估计","authors":"Giuseppe Boccignone, A. Marcelli, Paolo Napoletano, M. Ferraro","doi":"10.1109/ICIAP.2007.88","DOIUrl":null,"url":null,"abstract":"We present a probabilistic model for motion estimation in which motion characteristics are inferred on the basis of a finite mixture of motion models. The model is graphically represented in the form of a pairwise Markov random field network upon which a Loopy belief propagation algorithm is exploited to perform inference. Experiments on different video clips are presented and discussed.","PeriodicalId":118466,"journal":{"name":"14th International Conference on Image Analysis and Processing (ICIAP 2007)","volume":"177 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2007-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":"{\"title\":\"Motion Estimation via Belief Propagation\",\"authors\":\"Giuseppe Boccignone, A. Marcelli, Paolo Napoletano, M. Ferraro\",\"doi\":\"10.1109/ICIAP.2007.88\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We present a probabilistic model for motion estimation in which motion characteristics are inferred on the basis of a finite mixture of motion models. The model is graphically represented in the form of a pairwise Markov random field network upon which a Loopy belief propagation algorithm is exploited to perform inference. Experiments on different video clips are presented and discussed.\",\"PeriodicalId\":118466,\"journal\":{\"name\":\"14th International Conference on Image Analysis and Processing (ICIAP 2007)\",\"volume\":\"177 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2007-09-10\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"7\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"14th International Conference on Image Analysis and Processing (ICIAP 2007)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICIAP.2007.88\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"14th International Conference on Image Analysis and Processing (ICIAP 2007)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICIAP.2007.88","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 7

摘要

我们提出了一个运动估计的概率模型，其中运动特征是在有限混合运动模型的基础上推断出来的。该模型以成对马尔可夫随机场网络的形式图形化表示，在该网络上利用loop信念传播算法进行推理。在不同的视频片段上进行了实验并进行了讨论。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

Motion Estimation via Belief Propagation

We present a probabilistic model for motion estimation in which motion characteristics are inferred on the basis of a finite mixture of motion models. The model is graphically represented in the form of a pairwise Markov random field network upon which a Loopy belief propagation algorithm is exploited to perform inference. Experiments on different video clips are presented and discussed.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

14th International Conference on Image Analysis and Processing (ICIAP 2007)

自引率

0.00%

发文量

期刊最新文献

Real-time Gesture Recognition in Advanced Videocommunication Services Corner Displacement from Motion Blur A Method of Clustering Combination Applied to Satellite Image Analysis Sight enhancement through video fusion in a surveillance system Robust Iris Localization and Tracking based on Constrained Visual Fitting