Context-Aware Systems and Applications. First International Conference, ICCASA 2012, Ho Chi Minh City, Vietnam, November 26-27, 2012, Revised Selected Papers

Research Article

Object Detection and Tracking in Contourlet Domain

Download
340 downloads
  • @INPROCEEDINGS{10.1007/978-3-642-36642-0_20,
        author={Nguyen Thanh Binh and Tran Dien},
        title={Object Detection and Tracking in Contourlet Domain},
        proceedings={Context-Aware Systems and Applications. First International Conference, ICCASA 2012, Ho Chi Minh City, Vietnam, November 26-27, 2012, Revised Selected Papers},
        proceedings_a={ICCASA},
        year={2013},
        month={2},
        keywords={edge detection contourlet tranform cycle spinning object tracking},
        doi={10.1007/978-3-642-36642-0_20}
    }
    
  • Nguyen Thanh Binh
    Tran Dien
    Year: 2013
    Object Detection and Tracking in Contourlet Domain
    ICCASA
    Springer
    DOI: 10.1007/978-3-642-36642-0_20
Nguyen Thanh Binh1,*, Tran Dien1,*
  • 1: Ho Chi Minh City University of Technology
*Contact email: ntbinh@cse.hcmut.edu.vn, dientrananh@gmail.com

Abstract

This paper describes a method for the moving object detection and tracking in video sequences using contourlet transform. For the contourlet transform to be translation-invariant a 2D cycle spinning is implemented on subbands ∆  and ∆ . Cycle spinning for edge detection is implemented. The shape of object may change from this frame to other frame. The 3D moving object is combined two parts: a 2D shape change and 2D motion. The 2D motion of the object, we use the minimum Hausdorff distance from the model to the image to find where object moved to. With 2D shape change of the object, we use distance from the image to the transformed model to select set of image pixels of the next model. For performance evaluation, we compared the proposed method based on the contourlet transform using cycle spinning with the similar methods based on the complex wavelet transform and wavelet transform.