An adaptive bitrate method based on human visual characteristics

This paper proposes a novel framework for optimizing panoramic video transmission by integrating human visual system (HVS) characteristics, with a focus on non-uniform bitrate allocation. We introduce a viewport prediction algorithm, HVC360, that leverages visual attention mechanisms. The bitrate allocation problem is modeled as a Markov Decision Process (MDP), where the system dynamically adjusts bitrate distribution based on user viewport predictions and network conditions. Experimental results show that the proposed approach outperforms traditional methods, improving the comprehensive QoE index and HVC score by $9 \%$, and achieving an $81 \%$ allocation to the viewport, compared to $72 \%$ in conventional methods. The model demonstrates superior performance in long-term viewport prediction and efficient bandwidth utilization, offering a robust solution for next-generation immersive media transmission.