Semantic Classification and Retrieval Method of Arts and Crafts Images Based on Visual Feature Analysis

Content-based image retrieval technology overcomes the defect of subjective one-sidedness of text retrieval method and objectively describes the essential features of images. However, the semantic description of images in text mode is lost, which leads to the reduction of retrieval accuracy. To improve the performance of the retrieval system, image semantic description is used to improve the human-machine interface and interactive performance of the retrieval system. Then, the proposed visual feature-semantic classification algorithm is experimented to test its performance. The results on a random image set show that the proposed algorithm has better retrieval performance than traditional semantic classification algorithm and correlation sorting algorithm.