Information Technology Journal1812-56381812-5646Asian Network for Scientific Information10.3923/itj.2013.871.874YanshanLi WeiminLiu YujieCao 42013124The traditional moving object extraction method based on Gaussian
model has such defects as poor anti-noise performance, bad real-time performance.
Considering these shortcomings, this paper proposed a new moving object
extraction method for color video image based on regional kernel histogram.
The method first proposed the idea of kernel histogram description theory which
utilizing the kernel histogram to describe the area of video image. Then a new
metric function for measuring the kernel histogram model is proposed. According
to the features of measurement values, using the Gaussian mixture model and
the metric of kernel histogram model to build the model. At last, based on this
model, the moving object of video images is extracted. The experimental results
show that the algorithm have a better segmentation result and have the better
anti-noise and real-time performance compared with the traditional Gaussian
mixture model algorithm.]]>Cuevas, C., R. Mohedano and N. Garcia,20123731593161Min, J., H. Kim, J. Kim and I.S. Kweon,201248993995Sabirin, H. and M. Kim,201214657668Gilad, A.,19857384401Koga, T., K. Linuma, A. Hirano, Y. Iijima and T. Ishiguro,19811981pp: G5G3Ghanbari, M.,199038950953Stauffer, C. and W.E.L. Grimson,19991999pp: 246252Wayne Power, P. and J.A. Schoonees,20022002