The similar character of input video image can be generated by combining each components points after extracting the position of each component points using the ratio of face. By selecting the face area, in case of the upper half image, the neck and stained hair become change and accuracy of recognition of selected area becomes low, so in order to compensate the area of inaccuracy, we can find the shape of face, eyes and mouth applying the golden ratio and the character is reflected in the results. Using color informations of the image of the face, the TSL color can be analysed and shown the face recognition ratio of 1.7%. The components in each five input images (left and right eyes, eyebrows and mouth) by selecting the test results showed the performance of 88.3%. By analyzing the characteristics of the elements found in those features automatically generate the appropriate characters, and the rabbit character can be automatically generated from the animation story of 'Rabbit and Tortoise' made by flash tool.
展开▼