A Novel and Efficient Spatial–Temporal Saliency-Driven Integrated Video Compression

High-definition devices require the HD-videos for their proper utilization; however, it comprises several issues while transmission such as high computational complexity, long encoding time and restricted battery power. Moreover, several video compression algorithm has been introduced in past to solve the above-mentioned problem, however, due to the high-traffic video and low metrics of the existing algorithm, there is a requirement for an efficient algorithm. A major growth factor results in the contributions put forth towards video saliency, the existing methods perform saliency detection through a frame-wise approach that results in various challenges by incorporating an incoherent pixel-based saliency map detection that uses a spatio-temporal mechanism that utilizes frame-wise motion saliency with pixel-based temporal uniformity for diffusion purpose. This research develops an integrated video compression (IVC). At first, an effective and optimal spatio-temporal aware inter-frame and intra-frame-based saliency model is developed along with optimization modelling. Furthermore, two algorithms for designing a saliency map and optimized quantization for bitrate minimization. Performance analysis is carried out on a standard dataset; also comparison is carried out with existing state-of-art techniques to prove the model efficiency. IVC achieves better performance considering AUC, NCC, SIM and KL metrics. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd 2024.