Multi-stage Locally and Long-range Correlated Feature Fusion for Learned In-loop Filter in VVC
Multi-stage Locally and Long-range Correlated Feature Fusion for Learned In-loop Filter in VVC
复制标题
DOI：
10.1109/vcip56404.2022.10008834
复制
发表时间：
2022-12
期刊：
影响因子：
--
通讯作者：
B. Kathariya;Zhu Li;Hongtao Wang;G. V. D. Auwera
中科院分区：
文献类型：
--
作者：
B. Kathariya;Zhu Li;Hongtao Wang;G. V. D. Auwera
通用视频编码(VVC)/H266是目前最先进的视频编码标准，与之前的高效视频编码(HEVC)/H.26S相比，其编码效率有了显著提高。尽管如此，VVC也是基于块的视频编码技术，其中解码的图像包含压缩伪影。在VVC中，循环内滤波器用于抑制这些压缩伪影。本文利用卷积神经网络（CNN）来更好地抑制VVC上的压缩伪影。尽管如此，我们的方法在通过利用像素域的局部相关空间特征和离散余弦变换（DCT）域的远程相关光谱特征来获得更好的特征方面具有独特性。特别地，我们利用DCT变换后输入的cnn特征提取高频分量，并通过多阶段特征融合将远程相关性引入到空间cnn特征中。我们的实验结果表明，该方法在AI配置下对luma (Y)组件的平均Bjantegaard Delta （BD）比特率节省高达9.70%的显著编码改进。
Versatile Video Coding (VVC)/H.266 is currently the state-of-the-art video coding standard with significant improvement in coding efficiency over its predecessor High Efficiency Video Coding (HEVC)/H.26S. Nonetheless, VVC is also block-based video coding technology where decoded pictures contain compression artifacts. In VVC, in-loop filters serve to suppress these compression artifacts. In this paper, convolution neural network (CNN) is utilized to better facilitate the suppression of compression artifacts over VVC. Nonetheless, our approach has uniqueness in obtaining better features by exploiting locally correlated spatial features in the pixel domain as well as long-range correlated spectral features in the discrete cosine transform (DCT) domain. In particular, we utilized CNN-features from DCT transformed input to extract high-frequency components and induce long-range correlation into the spatial CNN-features by employing multi-stage feature fusion. Our experimental result shows that the proposed approach achieves significant coding improvements up to 9.70% on average Bjantegaard Delta (BD)-Bitrate savings under AI configurations for luma (Y) components.