Techniques Research Of Shape-Based Image Retrieval

In recent years visual information has become a major research area owing to the ever-increasing rate at which images are generated in many application fields. Visual information retrieval systems support retrieval by visual content by directly addressing image perceptual features such as color, shape and texture. But the shape-based image retrieval performs poor in the existing algorithm.In this paper, the background and the significance and the current progress of shape-based image retrieval are mentioned at first. Then the paper summarizes some classical algorithm of this field, two of them are stressed. One is edge orientation histogram, the other is chain code and Fourier descriptor. From the result of experiment, it can be seen that each of them has its own advantage and suits different kind of photos.Followed with the two methods, a system of shape-based image retrieval is built. Furthermore, this system is introduced in the paper.