Moving Objects Detection Algorithm Based on Visual Attention Mechanism and Gaussian Mixture Model
会议名称：《2011 International Conference on Signal,Image Processing and Applications（ICSIA 2011）》
会议日期：2011年
学科分类：08[工学] 080203[工学-机械设计及理论] 0802[工学-机械工程]
基 金：supported by the National Science Foundation of China(61065007) Foundation of Gansu Education,and the National Science Foundation of Gansu of China(0914ZTB098)
关 键 词：moving objects detection,gaussian mixture model,visual attention mechanism
摘 要：To deal with the disadvantage of noises and the problem that moving objects were replaced with background when the objects stay a long time in the scene,this paper introduced visual attention mechanism to improve the algorithm ***,space-time dynamic hybrid algorithm was adopted to deal with the moving images detected by Gaussian mixture model and the saliency map interested by visual attention ***,logicalandoperation was done between the significant moving objects of space-time dynamic hybrid mechanism and the saliency map region of visual attention mechanism to remove the isolate noises produced by Gaussian mixture *** simulations showed that the improved algorithm overcome the shortcoming effectively that moving objects were updated as background when the objects stay for a long time,and eliminated many isolate noises significantly.