Research on Scratch Detection of Old Movies

Abstract

Abstract Scratches are a continuous structural injury in old movies. Scratch pixel level segmentation algorithm still uses the general image segmentation algorithm, and the detection results are poor. According to the characteristics of scratches, a scratch model based on the Gaussian probability density function is proposed, a local adaptive one-dimensional Gaussian weighted segmentation algorithm is designed, and an invalid region filtering mechanism is introduced. Compared with several groups of experimental results, the local one-dimensional Gaussian weighted segmentation algorithm proposed in this paper can well preserve the integrity and continuity of scratches, and effectively suppress edge information, and the invalid region filtering mechanism can effectively reduce the false detection of scratches. Adding scratch trajectory can effectively reduce the missed detection rate of scratches, and the stability of the traditional algorithm using Kalman to predict scratch trajectory is poor. According to the similarity of multiple scratch trajectories, a calculation method of scratch trajectories based on U-net is designed. Experimental results show that this method can calculate scratch trajectory stably and effectively.