A Video Test Sequences Generation Way for Behavior Recognition Based on Cloud Model

Abstract:

Testing behavior recognition algorithm is a key technology of machine vision, tradition way to generate video test sequences has many defects, and a new way based on cloud model is putted forward. Moving object parameters such as moving types in image sequence were collected; these parameters were inputted into backward cloud generator. The quantitative representation of qualitative concept, expectation Ex,entropy En, super entropy He were gotten. These parameters were used to simulate the moving object behavior representation parameters. Ex, En, He were inputted into normal cloud generator, every moving object was designed as agent, agent can adjust behavior parameters by feeling environment and auto excitation. These behavior parameters were used to evaluate some classics algorithm. The experiment show that: the time cost of new way is 10% of tradition way, and two methods have same effect on testing algorithm recognition rate and algorithm recognition speed, and new way is superior to tradition way on testing algorithm robustness.

• Periodical: Advanced Materials Research (Volumes 591-593)
• Pages: 1276-1280
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.591-593.1276
• Online since: November 2012
• Authors: Min Li, Bin Wu
• Keywords: Behavior Recognition Algorithm Evaluation, Cloud Mode, Forward Cloud Generator, Moving Object Agent, Reverse Cloud Generator, Video Test Sequences Generation Method

References

[1] L.Wang, W.M.Hu, T.N. Tan. A Survey of Visual Analysis of Human Motion [J]. Journal of computers, vol.25 (2003) No.3,p.226-p.233. (In Chinese)

[2] Ahad MAR, Ogata T, Tan JK, et al. Motion recognition approach to solve overwriting in complex actions. [C] Proc. of the International Conference on Automatic Face and Gesture Recognition. (2008),pp.1-6.

DOI: 10.1109/afgr.2008.4813430

[3] Nguyen N T, Phung Q, Venkatesh S et al.Learning and detecting activities from movement trajectories using the hierarchical hidden Markov Model[C] .In Proc. of Computer Vision and Pattern Recognition. San Diego, (2005),pp.955-960.

DOI: 10.1109/cvpr.2005.203

[4] Mori G,Malik J, Recovering 3 D human body configurations using shape context[J] IEEE Trans on pattern Analysis and Machine Intelligence, Vol.28(2006) No.3,pp.1052-1062.

DOI: 10.1109/tpami.2006.149

[5] W.B. Wang. Object detection and recognition in Video Surveillance[M] Shanghai Jiao Tong University Publishing House,Vol.1( 2010),pp.153-157.(In Chinese)

[6] J.J. Huang,  P. F. Li. Multi target data association algorithm using cluster cloud model based c-means clustering [J].Journal of Shenzhen university science and engineering, Vol.01(2010) No.31,pp.11-15.(In Chinese)

[7] D.Y. Li .Uncertain Artificial Intelligence[M].Defense industry publishing house Vol.7(2005),pp.138-149.(In Chinese)

[8] Information on http://www.cbsr.ia.ac.cn