Researchers have developed NeR-SC, a novel neural representation framework specifically designed for screen content video compression. This framework builds upon the SNeRV backbone and introduces three key modules: a learnable color palette to model discrete color structures, a multi-gate dense fusion module for enhanced feature interaction, and an embedding-level frame skip strategy to bypass static frames, enabling real-time decoding. Experiments demonstrate that NeR-SC outperforms existing neural video representation methods and surpasses H.264 and H.265 at low bitrates, achieving competitive PSNR values on DSCVC and VCD datasets. AI
IMPACT This research advances neural video compression techniques, potentially leading to more efficient streaming and storage of screen content.
RANK_REASON The cluster contains two academic papers detailing new methods for neural video representation and compression.
AI-generated summary · Google Gemini · from 3 sources. How we write summaries →
