{"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/learning-content-weighted-deep-image","title":"Learning Content-Weighted Deep Image Compression","arxiv_id":"1904.00664","date":"2019-04-01","proceeding":null,"authors":["Mu Li","WangMeng Zuo","Shuhang Gu","Jane You","David Zhang"],"abstract":"Learning-based lossy image compression usually involves the joint\noptimization of rate-distortion performance. Most existing methods adopt\nspatially invariant bit length allocation and incorporate discrete entropy\napproximation to constrain compression rate. Nonetheless, the information\ncontent is spatially variant, where the regions with complex and salient\nstructures generally are more essential to image compression. Taking the\nspatial variation of image content into account, this paper presents a\ncontent-weighted encoder-decoder model, which involves an importance map subnet\nto produce the importance mask for locally adaptive bit rate allocation.\nConsequently, the summation of importance mask can thus be utilized as an\nalternative of entropy estimation for compression rate control. Furthermore,\nthe quantized representations of the learned code and importance map are still\nspatially dependent, which can be losslessly compressed using arithmetic\ncoding. To compress the codes effectively and efficiently, we propose a trimmed\nconvolutional network to predict the conditional probability of quantized\ncodes. Experiments show that the proposed method can produce visually much\nbetter results, and performs favorably in comparison with deep and traditional\nlossy image compression approaches.","url_abs":"http://arxiv.org/abs/1904.00664v1","url_pdf":"http://arxiv.org/pdf/1904.00664v1.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":"learning-content-weighted-deep-image","repo_url":"https://github.com/limuhit/CWIC","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"caffe2","reach":{"status":"ok"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"image-compression","task_name":"Image Compression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1904.00664","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}