영문내용
(English Abstract) |
Background subtraction is one of the very challenging tasks in image processing. The performance of the sample consensusbased background modeling consists of intensity correction, dynamically threshold and background sample updating, and postprocessing. Therefore, the performance depends on the aforementioned three processes. Most of the researchers tried to correct intensity, few worked on adaptive thresholding and dynamically background samples updating, and a very few researchers mentioned about post-processing operation. In this paper, we analyze the effect of post-processing in the background subtraction. Post-processing needs to be developed based on the characteristics of the raw foreground as our analysis suggests. We mention that post-processing plays one of the key roles to improve the performance of background subtraction. |
