Method for extracting target object from interactive video

The present invention discloses a method for extracting a target object in an interactive video, using a frame difference image acquiring three images before and after frame difference is determined as a rough estimate of the target object, then a rectangular binding region of the previous frame image coordinate position, determining a rectangular region of the current frame image included in the target object, and then construct an energy function for each frame image of the original video, accurate extraction target object, comprising the process of the rectangular area of ​​the target object in the determined frame image, full use of the coordinate position of the rectangular region of the previous frame image, so that even if a sudden stationary target object or background shaking, etc. are able to accurately determine the target object region of the current frame, the last of each image frame to determine a target area are established segmentation model , the target object can be extracted accurately by the method of FIG cut, thus greatly improving the extraction accuracy.