Adaptive Randomized Ensemble Tracking Using Appearance Variation and Occlusion Estimation
Tracking-by-detection methods have been widely studied with promising results. These methods usually train a classifier or a pool of classifiers in an online manner and use previous tracking results to generate a new training set for object appearance and update the current model to predict the obje...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
Hindawi Limited
2016-01-01
|
Series: | Mathematical Problems in Engineering |
Online Access: | http://dx.doi.org/10.1155/2016/1879489 |
id |
doaj-57d7b3978e88496d92ef6ed6d20fb96e |
---|---|
record_format |
Article |
spelling |
doaj-57d7b3978e88496d92ef6ed6d20fb96e2020-11-24T23:37:33ZengHindawi LimitedMathematical Problems in Engineering1024-123X1563-51472016-01-01201610.1155/2016/18794891879489Adaptive Randomized Ensemble Tracking Using Appearance Variation and Occlusion EstimationWeisheng Li0Yanjun Lin1Chongqing Key Laboratory of Computational Intelligence, Chongqing University of Posts and Telecommunications, Chongqing 400065, ChinaChongqing Key Laboratory of Computational Intelligence, Chongqing University of Posts and Telecommunications, Chongqing 400065, ChinaTracking-by-detection methods have been widely studied with promising results. These methods usually train a classifier or a pool of classifiers in an online manner and use previous tracking results to generate a new training set for object appearance and update the current model to predict the object location in subsequent frames. However, the updating process may easily cause drifting in terms of appearance variation and occlusion. The previous methods for updating the classifier(s) decided whether or not to update the classifier(s) by a fixed learning rate parameter in all scenarios. The learning rate parameter has a great influence on the tracker’s performance and should be dynamically adjusted according to the change of scene during tracking. In this paper, we propose a novel method to model the time-varying appearance of an object that takes appearance variation and occlusion of local patches into consideration. In contrast with the existing methods, the learning rate for updating classifier ensembles adaptively is adjusted by estimating the appearance variation with sparse optical flow and the possible occlusion of the object between consecutive frames. Experiments and evaluations on some challenging video sequences have been done and the results demonstrate that the proposed method is more robust against appearance variation and occlusion than those state-of-the-art approaches.http://dx.doi.org/10.1155/2016/1879489 |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Weisheng Li Yanjun Lin |
spellingShingle |
Weisheng Li Yanjun Lin Adaptive Randomized Ensemble Tracking Using Appearance Variation and Occlusion Estimation Mathematical Problems in Engineering |
author_facet |
Weisheng Li Yanjun Lin |
author_sort |
Weisheng Li |
title |
Adaptive Randomized Ensemble Tracking Using Appearance Variation and Occlusion Estimation |
title_short |
Adaptive Randomized Ensemble Tracking Using Appearance Variation and Occlusion Estimation |
title_full |
Adaptive Randomized Ensemble Tracking Using Appearance Variation and Occlusion Estimation |
title_fullStr |
Adaptive Randomized Ensemble Tracking Using Appearance Variation and Occlusion Estimation |
title_full_unstemmed |
Adaptive Randomized Ensemble Tracking Using Appearance Variation and Occlusion Estimation |
title_sort |
adaptive randomized ensemble tracking using appearance variation and occlusion estimation |
publisher |
Hindawi Limited |
series |
Mathematical Problems in Engineering |
issn |
1024-123X 1563-5147 |
publishDate |
2016-01-01 |
description |
Tracking-by-detection methods have been widely studied with promising results. These methods usually train a classifier or a pool of classifiers in an online manner and use previous tracking results to generate a new training set for object appearance and update the current model to predict the object location in subsequent frames. However, the updating process may easily cause drifting in terms of appearance variation and occlusion. The previous methods for updating the classifier(s) decided whether or not to update the classifier(s) by a fixed learning rate parameter in all scenarios. The learning rate parameter has a great influence on the tracker’s performance and should be dynamically adjusted according to the change of scene during tracking. In this paper, we propose a novel method to model the time-varying appearance of an object that takes appearance variation and occlusion of local patches into consideration. In contrast with the existing methods, the learning rate for updating classifier ensembles adaptively is adjusted by estimating the appearance variation with sparse optical flow and the possible occlusion of the object between consecutive frames. Experiments and evaluations on some challenging video sequences have been done and the results demonstrate that the proposed method is more robust against appearance variation and occlusion than those state-of-the-art approaches. |
url |
http://dx.doi.org/10.1155/2016/1879489 |
work_keys_str_mv |
AT weishengli adaptiverandomizedensembletrackingusingappearancevariationandocclusionestimation AT yanjunlin adaptiverandomizedensembletrackingusingappearancevariationandocclusionestimation |
_version_ |
1725519609847611392 |