The research of a video segmentation algorithm based on image fusion in the wavelet domain

Fan Yang; Jing Li; Shun-hua Xu; Guo-feng Pan

doi:10.1117/12.864536

12 October 2010 The research of a video segmentation algorithm based on image fusion in the wavelet domain

Fan Yang, Jing Li, Shun-hua Xu, Guo-feng Pan

Proceedings Volume 7659, 5th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Smart Structures and Materials in Manufacturing and Testing; 765918 (2010) https://doi.org/10.1117/12.864536
Event: 5th International Symposium on Advanced Optical Manufacturing and Testing Technologies, 2010, Dalian, China

Abstract

For the shortcoming of a large number of computational complexity and time consuming of traditional Spatio-Temporal Fusion Segmentation Algorithm, an automatic video object segmentation algorithm based on the Spatio-Temporal Fusion in wavelet domain is proposed. The temporal segmentation in wavelet domain, firstly, wavelet transforms adjacent 3 frame image and uses a sub-band bi-directional difference method, then makes the two differential images fuse. In the low-frequency sub-band, choose options for integration strategies or a weighted average fusion strategy by setting the matching threshold; In the high-frequency sub-band, using the pixel that has the larger absolute value Strategy, which can enhance the edge information of the image and derive more movable information. Secondly, the spatial segmentation performs the improved watershed segmentation algorithm with the gradient and combination of internal and external constraints watershed segmentation algorithm in wavelet domain to resolve the issue of over-segmentation. Meanwhile, mark and extract sub-block segmentation region as the spatial segmentation template of the wavelet domain. The spatio-temporal fusion segmentation template in wavelet sub-band can be derived by mapping the spatial segmentation to the temporal one, statistics the probability of '1', which is belong to video object region (compose of '0' and '1') and then making them fusion in wavelet domain. Then after 'and' operation with the current frame and wavelet reverse transform, the outcome of the final segmentation is derived. The experimental results have shown that the method has a good result, i.e. the rooms' numbers of the improved spatial segmentation are 2~3 times less than traditional method, the processing speed is 8~9 times advanced averagely for each frame and the computational complexity of template operation is greatly reduced.

Citation Download Citation

Fan Yang, Jing Li, Shun-hua Xu, and Guo-feng Pan "The research of a video segmentation algorithm based on image fusion in the wavelet domain", Proc. SPIE 7659, 5th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Smart Structures and Materials in Manufacturing and Testing, 765918 (12 October 2010); https://doi.org/10.1117/12.864536

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available