Knowledge Commons of Institute of Automation,CAS
Clustering Ensemble Tracking | |
Guibo Zhu; Jinqiao Wang; Hanqing Lu | |
2014 | |
会议名称 | ACCV2014 |
会议录名称 | Asian Conference on Computer Vision (ACCV) |
页码 | 382-396 |
会议日期 | 2014 |
会议地点 | Singapore |
摘要 | A key problem in visual tracking is how to handle the ambiguity in decision to locate the object effectively using the target appearance model with online update. We address this problem by incorporating sequential clustering and ensemble methods into the tracking system.In this paper, clustering is used for mining the potential historical structure in the parameter space and feature space. Then we fuse multiple weak hypotheses to construct a strong ensemble learner for object tracking. Different from previous methods for updating classifier ensemble in a fixed weak classifier pool frame-to-frame, the proposed ensemble method is taking three weak hypotheses into consideration: spatial object-part view, parameter space view, and feature space view. Specially, spatial object-part view represents the object by a collection of part models that are spatially related (e.g. tree-structure). Meanwhile, analyzing the latent group structure in the parameter space and feature space is essential to take full advantage of the historical data in the tracking process. Therefore, we propose a novel ensemble algorithm that fuses object-part predictor, parameter clustered predictors and feature clustered predictors together. Furthermore, the weights of different views are updated by the relative consistency between weak predictors and final ensemble tracker. The formulation is tested in a tracking-by-detection implementation. Extensive comparing experiments on challenging video sequences demonstrate the robustness and effectiveness of the proposed method. |
关键词 | Visual Tracking Online Clustering Part Model |
收录类别 | EI |
语种 | 英语 |
文献类型 | 会议论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/4698 |
专题 | 紫东太初大模型研究中心_图像与视频分析 |
通讯作者 | Jinqiao Wang |
推荐引用方式 GB/T 7714 | Guibo Zhu,Jinqiao Wang,Hanqing Lu. Clustering Ensemble Tracking[C],2014:382-396. |
条目包含的文件 | ||||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
Clustering Ensemble (708KB) | 会议论文 | 开放获取 | CC BY-NC-SA | 浏览 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论