率失真优化 | ViLab https://vilab.team/tag/%E7%8E%87%E5%A4%B1%E7%9C%9F%E4%BC%98%E5%8C%96/ 率失真优化 Hugo Blox Builder (https://hugoblox.com)en-usTue, 18 Jul 2023 00:00:00 +0000 https://vilab.team/media/icon_hu2896232876136423579.png 率失真优化 https://vilab.team/tag/%E7%8E%87%E5%A4%B1%E7%9C%9F%E4%BC%98%E5%8C%96/ Learned rate-distortion cost prediction for ultrafast screen content intra coding https://vilab.team/publication/learned-rate-distortion-cost-prediction-for-ultrafast-screen/ Tue, 18 Jul 2023 00:00:00 +0000 https://vilab.team/publication/learned-rate-distortion-cost-prediction-for-ultrafast-screen/ <p>本文面向实时屏幕内容通信中的超快编码需求，提出基于学习的率失真代价预测方法。该方法不再通过实际编码计算RD代价，而是构建神经网络分别预测帧内预测、调色板及普通IBC模式的RD代价，并对IBC merge模式结合运动补偿与线性回归进行预测。利用预测结果生成分区-模式映射集，从而显著降低H.265/HEVC SCC扩展的编码复杂度，在保持编码效率的同时实现超快屏幕内容帧内编码。</p>