ADAPTIVE MAPPING PARAMETER DETECTION METHOD

摘要

PURPOSE: An adaptive transform parameter detecting method is provided to be capable of reducing an excessive calculation amount and minimizing the loss of a screen quality. CONSTITUTION: A difference image is looked for between a current image and a previous image(S11). There is carried out labeling with regard to an object partition and a moving object(S12). A temporal co-occurrence matrix is calculated with regard to each object(S13). A ratio(K) of each object is calculated according to a predetermined threshold value(Th1)(S14). A difference image reduced into a 1/4 size is looked for by a hierarchical method(S15). A K value of a corresponding object is compared with a threshold value(Th2)(S16). If the K value is less than the Th2 value, a 6-parameter is detected from an original image of a low resolution using a difference image of a low resolution(S17) and an image of a low resolution is synthesized(S18). A parameter is verified by comparing the 6-parameter with the synthesized image(S19). If successfully verified, boundary information between the 6-parameter and the object is transferred(S20). If failed, approximation information and boundary information are transferred(S25).