A Multi-Modal Video Data Dimensionality and Replication Reduction on Distributed Architecture

Abstract

With the growth in the multimedia multi-modal data sharing and improvement in the computational capabilities, the demand for the research in this domain is also increasing. In the recent past a good number of research attempts can be observed in order to make the video-based data processing for making faster searching possible. Nevertheless, most of the parallel research attempts are criticised for two reasons. The extraction of the key frame for reduction in the data size, which can improve the processing time and the complexity of the modal. Also, the replication of the data must be reduced in order to further reduce the searching time. This work proposes two novel methods for reduction in the keyframe size and for replication control. The keyframe extraction method is realized using the global and local thresholding methods by deploying the adjustment of correction factors as per the machine learning methods and the replication control is realized using the histogram method. During the keyframe extraction process, the proposed algorithm has demonstrated reduction in the keyframevs actual frame ratio from 0.8852 to 0.0080 as mean, which is nearly 95% of reduction and during the replication control, the proposed algorithm has demonstrated nearly 100% accuracy. Nonetheless, the validation of any key frame extraction process completely relies on the information extraction compared with the actual video data. The information loss must be reduced to 0% in order to establish the believe that the key frame extraction process, which is proposed is 100% lossless. Thus, this work also verifies the obtained keyframes with information extraction process and obtained 100% information retrieval for the same purposes. The primary outcome of this work is to reduce the time and space complexity of the multi-modal distributed video data processing for making the video-based search engines a more efficient method for content retrieval and delivery.