{"title":"Poisson Kalman Particle Filtering for Tracking Centrosomes in Low-Light 3-D Confocal Image Sequences","authors":"Hugh Gribben, P. Miller, Jianguo Zhang, M. Browne","doi":"10.1109/IMVIP.2009.22","DOIUrl":null,"url":null,"abstract":"An automatic tracker is developed, which is capable of tracking intra-cellular features in living cells from 3-D confocal image sequences corrupted by noise. The proposed approach takes a Poisson MAP-MRF classification as an initial stage to detect objects. These are then used to update the multiple target locations generated by 3D Poisson Kalman Particle filters (PKPF). A probabilistic nearest neighbour search strategy for object association is developed to produce improved prediction of target locations. Our approach is tested in real 3D confocal image sequences with challenging illumination conditions. Results show that our Poisson Kalman particle filter approach obtains very promising results and outperforms three other tracking approaches.","PeriodicalId":179564,"journal":{"name":"2009 13th International Machine Vision and Image Processing Conference","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2009-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2009 13th International Machine Vision and Image Processing Conference","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IMVIP.2009.22","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
An automatic tracker is developed, which is capable of tracking intra-cellular features in living cells from 3-D confocal image sequences corrupted by noise. The proposed approach takes a Poisson MAP-MRF classification as an initial stage to detect objects. These are then used to update the multiple target locations generated by 3D Poisson Kalman Particle filters (PKPF). A probabilistic nearest neighbour search strategy for object association is developed to produce improved prediction of target locations. Our approach is tested in real 3D confocal image sequences with challenging illumination conditions. Results show that our Poisson Kalman particle filter approach obtains very promising results and outperforms three other tracking approaches.