Learned Image Compression Using Cross-Component Attention Mechanism

作     者：Duan, Wenhong Chang, Zheng Jia, Chuanmin Wang, Shanshe Ma, Siwei Song, Li Gao, Wen 

作者机构：Shanghai Jiao Tong Univ Dept Comp Sci & Engn Shanghai 200240 Peoples R China Shanghai Jiao Tong Univ AI Inst Shanghai 200240 Peoples R China Chinese Acad Sci Inst Comp Technol Beijing 100190 Peoples R China Peking Univ Wangxuan Inst Comp Technol WICT Beijing 100871 Peoples R China Peking Univ Natl Engn Res Ctr Visual Technol Sch Comp Sci Beijing 100871 Peoples R China Shanghai Jiao Tong Univ Inst Image Commun & Network Engn AI Inst Shanghai 200240 Peoples R China 

出 版 物：《IEEE TRANSACTIONS ON IMAGE PROCESSING》 (IEEE Trans Image Process)

年 卷 期：2023年第32卷

页      面：5478-5493页

核心收录：

学科分类：0808[工学-电气工程] 08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：National Natural Science Foundation of China [62025101, 62101007] Fundamental Research Funds for the Central Universities Young Elite Scientist Sponsorship Program by the Beijing Association for Science and Technology (BAST) [BYSS2022019] Wen-Tsun Wu Honorary Doctoral Scholarship AI Institute Shanghai Jiao Tong University 

主　　题：Image coding Context modeling Transforms Decoding Standards Image reconstruction Transform coding Image compression cross-component information-guided unit attention mechanism information-preserving 

摘      要：Learned image compression methods have achieved satisfactory results in recent years. However, existing methods are typically designed for RGB format, which are not suitable for YUV420 format due to the variance of different formats. In this paper, we propose an information-guided compression framework using cross-component attention mechanism, which can achieve efficient image compression in YUV420 format. Specifically, we design a dual-branch advanced information-preserving module (AIPM) based on the information-guided unit (IGU) and attention mechanism. On the one hand, the dual-branch architecture can prevent changes in original data distribution and avoid information disturbance between different components. The feature attention block (FAB) can preserve the important information. On the other hand, IGU can efficiently utilize the correlations between Y and UV components, which can further preserve the information of UV by the guidance of Y. Furthermore, we design an adaptive cross-channel enhancement module (ACEM) to reconstruct the details by utilizing the relations from different components, which makes use of the reconstructed Y as the textural and structural guidance for UV components. Extensive experiments show that the proposed framework can achieve the state-of-the-art performance in image compression for YUV420 format. More importantly, the proposed framework outperforms Versatile Video Coding (VVC) with 8.37% BD-rate reduction on common test conditions (CTC) sequences on average. In addition, we propose a quantization scheme for context model without model retraining, which can overcome the cross-platform decoding error caused by the floating-point operations in context model and provide a reference approach for the application of neural codec on different platforms.