Massive Real-Time Data Mining Algorithm for a Multimedia Database

The traditional classification mining algorithm for massive data has a complicated calculation process, poor real-time performance and low accuracy. This paper presents a real-time data mining algorithm for a multimedia database. The wavelet de-noising method is used to de-noise the data in the multimedia database to reduce the interference of the background area to the feature extraction of the multimedia data. The effective area extraction algorithm based on the center of mass is used to extract the effective area and is combined with the SIFT algorithm and LBP algorithm to extract the data features from the multimedia database. Experimental results show that the algorithm is accurate, reliable, has a high real-time mining ability and a practical ability. © 2022 CRL Publishing Ltd