{"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/an-end-to-end-compression-framework-based-on","title":"An End-to-End Compression Framework Based on Convolutional Neural Networks","arxiv_id":"1708.00838","date":"2017-08-02","proceeding":null,"authors":["Feng Jiang","Wen Tao","Shaohui Liu","Jie Ren","Xun Guo","Debin Zhao"],"abstract":"Deep learning, e.g., convolutional neural networks (CNNs), has achieved great\nsuccess in image processing and computer vision especially in high level vision\napplications such as recognition and understanding. However, it is rarely used\nto solve low-level vision problems such as image compression studied in this\npaper. Here, we move forward a step and propose a novel compression framework\nbased on CNNs. To achieve high-quality image compression at low bit rates, two\nCNNs are seamlessly integrated into an end-to-end compression framework. The\nfirst CNN, named compact convolutional neural network (ComCNN), learns an\noptimal compact representation from an input image, which preserves the\nstructural information and is then encoded using an image codec (e.g., JPEG,\nJPEG2000 or BPG). The second CNN, named reconstruction convolutional neural\nnetwork (RecCNN), is used to reconstruct the decoded image with high-quality in\nthe decoding end. To make two CNNs effectively collaborate, we develop a\nunified end-to-end learning algorithm to simultaneously learn ComCNN and\nRecCNN, which facilitates the accurate reconstruction of the decoded image\nusing RecCNN. Such a design also makes the proposed compression framework\ncompatible with existing image coding standards. Experimental results validate\nthat the proposed compression framework greatly outperforms several compression\nframeworks that use existing image coding standards with state-of-the-art\ndeblocking or denoising post-processing methods.","url_abs":"http://arxiv.org/abs/1708.00838v1","url_pdf":"http://arxiv.org/pdf/1708.00838v1.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":"an-end-to-end-compression-framework-based-on","repo_url":"https://github.com/compression-framework/compression_framwork_for_tesing","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null},{"paper_slug":"an-end-to-end-compression-framework-based-on","repo_url":"https://github.com/TristansCloud/YellowstonesVegitiation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"an-end-to-end-compression-framework-based-on","repo_url":"https://github.com/kunalrdeshmukh/End-to-end-compression","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"an-end-to-end-compression-framework-based-on","repo_url":"https://github.com/piyushpandey615/GAN-based-Image-Compressor-System","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"an-end-to-end-compression-framework-based-on","repo_url":"https://github.com/ppooiiuuyh/ComRecCNN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"image-compression","task_name":"Image Compression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1708.00838","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}