Video data retrieval is one of the prime focus for recent research trends for information processing and management domain. The popularity of the video data for various purposes is increasing and the set of users developing the video contents for many purposes are also increasing. As a result, many of video content sharing and storing services are also gaining popularity. In this situation, the management of the video data storage is moved on to the cloud computing environment. Due to the replicated and distributed job processing architecture of cloud computing, the video data is shared over storage clusters and many of the time, instead of replicating the data over various clusters, the cloud service providers distribute the data using percentage split to reduce storage cost. This not only reduces the storage costs, but also decreases the information processing time. This increases the complexity of the video search engines. Due to the percentage split, mining data can cause duplication issues and increases the time complexity to search the data. A number of research attempts have aimed to solve this bottleneck, but the recent outcomes are criticised for being highly time complex. Hence, this work proposes a novel strategy to retrieve the video contents. The present work proposing a two-fold algorithm for video data retrieval. Firstly, the problem of local thresholding during retrieval is eliminated with the introduction of adaptive and adjustable thresholding for the cluster specific video content search results and secondly, the time complexity is reduced with the introduction of local Knowledge Discovery Factor (KDF). The algorithm demonstrates a nearly 99% accuracy in the video data retrieval process with nearly 20% improvement for time complexity.
Volume 11 | Issue 5