Cover Page

Video Object Wavelet (VOW) Coder

Figure 1. Sample Video Object

We have worked on the compression of arbitrary shape object. We have proposed a new wavelet transform for the arbitrary shape object, and a partial bitplane embedded coder for the compression of the arbitrary shape object. We call the coder the video object wavelet (VOW) coder. Experimental results show that VOW significantly outperforms MPEG-4 in still object coding, and achieves a comparable performance in video object coding in terms of PSNR. Moreover, the VOW decoded object looks better subjectively, with less annoying blocking artifacts than that of MPEG-4.

For details of the work, please refer to the following paper.

J. Li and S. Lei, "Arbitrary shape wavelet transform with phase alignment", 1998 IEEE International Conference on Image Processing, volume 3, pp. 683-687, Chicago, IL, Oct. 1998. [The arbitrary shape wavelet transform algorithm, the algorithm is independently developed by S. Li and W. Li at Lehigh University.]
G. Xing, J. Li, S. Li and Y. Zhang, "Arbitrarily shaped video object coding by wavelet", IEEE International Symposium for Circuit and Systems (ISCAS 2000), Geneva, Switzerland, May. 2000. [Lifting implementation of arbitrary shape wavelet transform with phase alignment, and partial bitplane coding of the wavelet coefficients. ]
G. Xing, J. Li, S. Li and Y. Zhang, "Arbitrarily shaped video object coding by wavelet", to be published by IEEE Trans. On Circuits and Systems for Video Technology. [Journal version of the above paper.]