LAUSR

Aiming at the problems of poor precision and recall, long retrieval time and high energy consumption in current video image indexing methods, a local feature indexing method for multimedia video based on intelligent soft computing is proposed. Video image is segmented by maximum entropy threshold method. Based on the result of segmentation, features are clustered in two-dimensional space. Each video image is divided into several feature groups. Unified descriptors are generated for each feature group. The descriptors of each feature group are coded by binary coding. The similarity between index items and video images in database is calculated, and local feature indexing of media video is realized by looking up tables. The experimental results show that the method has high index precision and recall, low energy consumption and real-time performance. The proposed method has excellent performance and robustness.