A Modified Frame Difference Method Using Correlation Coefficient for Background Subtraction

Background subtraction is one of the most important step in video surveillance which is used in a number of real life applications such as surveillance, human machine interaction, optical motion capture and intelligent visual observation of animals, insects. Background subtraction is one of the preliminary stages which are used to differentiate the foreground objects from the relatively stationary background. Normally a pixel is considered as foreground if its value is greater than its value in the reference image. Hence, every pixel has to be compared to find the foreground and background pixel. This paper presents a technique which improves the frame difference method by first classifying the blocks in the frame as background and others using correlation coefficient. Further refinement is performed by performing pixel-level classification on blocks which are not considered as background. Experiments are conducted on standard data-sets and the performance measures shows good results in some critical conditions.