透過您的圖書館登入
IP:3.147.42.168
  • 學位論文

加入線上模板資料庫建構之三維物體追蹤平均位移演算法

3D Object Mean-shift Tracking Algorithm with Online Template Database Construction

指導教授 : 胡竹生

摘要


本論文提出了一套建構在三維物體追蹤演算法下,可用於複雜背景與移動場景的目標模型更新的方法。影像追蹤演算法採用三維平均位移演算法來計算目標物於空間中的位移,並利用主成份分析法來估測物體於三維空間中的大小與旋轉,由於加入了三維空間資訊以及新的目標模型更新方式,因此在複雜背景下以及目標物外觀與顏色產生較大變化時,會比使用其他的目標模型更新方法具有更穩健的追蹤效果。另外提出加入線上模板資料庫建構的方法,資料庫建構的準則以及資料庫內的資料選取將是本論文的另一個重點,這會使得追蹤目標物的過程中,對於目標模型的外觀及顏色變化有更迅速及適應性更高的效果,當目標物因旋轉而導致外觀顏色變化較為劇烈時,比起一般其他直接對目標顏色機率模型做更新的方法,準確性與穩定性來得更高。

並列摘要


In this thesis, a target model updating method based on 3-D mean-shift tracking algorithm is proposed. The algorithm is aimed at reducing the influence of complex and dynamic background. It uses principal component analysis to estimate the size and rotation of the target in 3-Dimensions. With the 3-D spatial information and new target model updating method, this algorithm outperforms other target model updating methods such as traditional Mean-Shift and 3-D Mean-Shift. The method consists of a set of rules to construct online template database. These results in a better and faster tracking process. In particular, when target rotates drastically, it gives a more accurate and stable result compared to other methods that only update the target model.

並列關鍵字

target model update online database 3D mean-shift algorithm tracking

參考文獻


[2] D. Comaniciu, V. Ramesh, and P. Meer, “Kernel-based object tracking,” IEEE Trans. On Pattern Analysis and Machine Intelligence, 25(5):564-577, May 2003.
[3] Q. Zhao and H. Tao, “Object tracking using color correlogram,” Proc. IEEE Conf. on Computer Vision and Pattern Recognition, 2005.
[4] S. Birchfield and S. Rangarajan, “Spatiograms versus histograms for region-based tracking,” Proc. IEEE Conf. on Computer Vision and Pattern Recognition, 2005.
[5] H. Zhang, Z. Huang, W. Huang, and L. Li, “Kernel-based method for tracking objects with rotation and translation,” Proc. International Conf. on Pattern Recognition, 2004.
[6] Z. Zivkovic and B. Krose, “An em-like algorithm for color-histogram-based object tracking,” Proc. IEEE Conf. on Computer Vision and Pattern Recognition, 2004.

延伸閱讀