{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/reducing-complexity-of-hevc-a-deep-learning","title":"Reducing Complexity of HEVC: A Deep Learning Approach","arxiv_id":"1710.01218","date":"2017-09-19","proceeding":null,"authors":["Mai Xu","Tianyi Li","Zulin Wang","Xin Deng","Ren Yang","Zhenyu Guan"],"abstract":"High Efficiency Video Coding (HEVC) significantly reduces bit-rates over the\nproceeding H.264 standard but at the expense of extremely high encoding\ncomplexity. In HEVC, the quad-tree partition of coding unit (CU) consumes a\nlarge proportion of the HEVC encoding complexity, due to the bruteforce search\nfor rate-distortion optimization (RDO). Therefore, this paper proposes a deep\nlearning approach to predict the CU partition for reducing the HEVC complexity\nat both intra- and inter-modes, which is based on convolutional neural network\n(CNN) and long- and short-term memory (LSTM) network. First, we establish a\nlarge-scale database including substantial CU partition data for HEVC intra-\nand inter-modes. This enables deep learning on the CU partition. Second, we\nrepresent the CU partition of an entire coding tree unit (CTU) in the form of a\nhierarchical CU partition map (HCPM). Then, we propose an early-terminated\nhierarchical CNN (ETH-CNN) for learning to predict the HCPM. Consequently, the\nencoding complexity of intra-mode HEVC can be drastically reduced by replacing\nthe brute-force search with ETH-CNN to decide the CU partition. Third, an\nearly-terminated hierarchical LSTM (ETH-LSTM) is proposed to learn the temporal\ncorrelation of the CU partition. Then, we combine ETH-LSTM and ETH-CNN to\npredict the CU partition for reducing the HEVC complexity for inter-mode.\nFinally, experimental results show that our approach outperforms other\nstate-of-the-art approaches in reducing the HEVC complexity at both intra- and\ninter-modes.","url_abs":"http://arxiv.org/abs/1710.01218v3","url_pdf":"http://arxiv.org/pdf/1710.01218v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"reducing-complexity-of-hevc-a-deep-learning","repo_url":"https://github.com/HEVC-Projects/CPH","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"deep-learning","task_name":"Deep Learning"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[{"slug":"cph","name":"CPH","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}