Research On Feature Ex-traction And Recognition For Ani-mation Resource

Globally, the animation industry has become a huge industry. However, in China's animation industry is still an emerging industry, but is developing rapidly. In the ani-mation industry there is an important problem which is how to finish feature extraction and classify recognition efficiently. Therefore, the motivation of this paper is to solve this problem by use of digital image processing technology.Animation material is mainly in the form of images. Paper-cut image, as a typical anime material, is different from the ordinary image on the features and characteristics of human visual image. Therefore, according to the characteristics of paper-cut images, we studied feature extraction and recognition technology of paper-cut image, which can be used to solve the above proposed problems.Firstly, the paper provides feature extraction method of Paper-cut image based on space information by combining Spatial Pyramid Matching and Contextualizing His-togram, which overcome the restriction problem of Spatial Pyramid Matching or Con-textualizing Histogram. In addition, traditional image recognition methods cannot in-crease accuracy without space information. This paper proposes an image recognition method based on feature combination and selection by using AdaBoost method which can solve this problem. Experiments show that the proposed method can achieve better performance.