Foreground moving object detection plays a critical role in video surveillance systems. Local Binary Pattern (LBP) based fuzzy distance segmentation is proposed for object foreground detection which uses modified training rate that deals with a change in the foreground of different rates. The extracted features are trained by using hybrid self-organizing map (SOM) based artificial bee colony (ABC) optimization. This approach deals with dynamic background and the illumination changes effectively. In this fuzzy c_means based segmentation effectively segments the moving object for foreground detection. Then, the feature from the foreground is extracted using the LBP feature extraction technique. Thus, foreground image can be further classified for identifying the object under moving using SOM with optimization. The proposed work is simulated in the working platform of Matlab, and the performance is analyzed by comparing with the existing techniques. Experimental results declare that the proposed method accomplishes more changes in the precision over existing techniques.

Username
Password