Deep Learning-Enabled Efficient Storage and Retrieval of Video Streams in the Cloud
Mahmoud Darwich, Taghreed Alghamdi, Magdy Bayoumi · 2023
In recent years, the rapid growth of video data has presented challenges for efficient storage and retrieval in cloud-based systems. This research proposes a novel approach that utilizes machine learning techniques to address these challenges. The objective is to develop a robust framework that optimizes storage utilization, enhances retrieval efficiency, and maintains video quality with minimal latency. Machine learning algorithms are employed for tasks such as video compression, content analysis, and metadata extraction, enabling effective storage and retrieval mechanisms. The framework leverages advanced video analytics algorithms to extract relevant features, incorporating them into the storage and retrieval process for efficient indexing and organization of video content. The framework adapts to changing workloads and dynamically allocates resources for optimal performance in cloud environments. Extensive experiments and comparisons with existing methods validate the proposed framework, demonstrating significant improvements in storage utilization, retrieval performance, and scalability. Specifically, our framework achieves an average improvement of 25% in storage efficiency compared to another method. It also demonstrates faster retrieval speeds, with an improvement of approximately 10-20%, and achieves higher video quality ratings, with an improvement of approximately 10-20% compared to existing methods.