Background Separation in Motion Picture Using Presumptive Eyes Position and Different Image

摘要

The paper used the method using the information of colors by converting RGB information, the color image information to HSI model and YC_bC_r model and the presumptive eyes position, as the detection method of the region of face and its minutiae. From the sample image, the study suggested the method to separate and extract a subject from its background even in the sample images where there are colors similar to a given skin colors and to replace the original background with another background. Then, by using the only 4.8% of the overall region, the initial region of interest to comprehend filtering and a moving object was created, which demonstrates that based on the overall region, the only 13% may be used to track an actually moving object.