Motion estimation using coarse to fine based on wavelet multiresolution for inbetweening in old animation films

Animation film is much demand entertainment. If the old animation films have audio-visual quality as desired modern audience then the audience would be interested in the film. The old animation films have several kinds of damage such as intensity flicker, noise, blotches and movement of objects in the video is less smooth (jerky). Inbetweening is the process of generating intermediate frames between two images to give the appearance that the first image evolves smoothly into the second image. Motion estimation is required for inbetweening. Animation film object of this study is the result of manual hand drawing and graylevel color so it requires the determination of motion vector estimation is precise and quick. In this paper we propose a quality enhancement method of the object motion of old animation film using inbetweening. The inbetweening method which is proposed uses object interpolation approach based on motion vector. Motion estimation is proposed using coarse to fine based on wavelet multiresolution, while the function of matching using MAD. The first step is eliminating the defects of old animation film in order to obtain the accurate motion vector estimation. Then, motion estimation to find motion vector and refinement that. Then, Segment a frame into objects and background. After that, the inbetweening process is executed by using object interpolation approach. In the experiment using the old animation films are produced in the 1920s. Experimental results show that our proposed method have average PSNR 23.46 dB for Popeye sequence, 31.73 for Charlie Chaplin sequence, and 30.82 for felix sequence.