{"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/virtual-codec-supervised-re-sampling-network","title":"Virtual Codec Supervised Re-Sampling Network for Image Compression","arxiv_id":"1806.08514","date":"2018-06-22","proceeding":null,"authors":["Lijun Zhao","Huihui Bai","Anhong Wang","Yao Zhao"],"abstract":"In this paper, we propose an image re-sampling compression method by learning\nvirtual codec network (VCN) to resolve the non-differentiable problem of\nquantization function for image compression. Here, the image re-sampling not\nonly refers to image full-resolution re-sampling but also low-resolution\nre-sampling. We generalize this method for standard-compliant image compression\n(SCIC) framework and deep neural networks based compression (DNNC) framework.\nSpecifically, an input image is measured by re-sampling network (RSN) network\nto get re-sampled vectors. Then, these vectors are directly quantized in the\nfeature space in SCIC, or discrete cosine transform coefficients of these\nvectors are quantized to further improve coding efficiency in DNNC. At the\nencoder, the quantized vectors or coefficients are losslessly compressed by\narithmetic coding. At the receiver, the decoded vectors are utilized to restore\ninput image by image decoder network (IDN). In order to train RSN network and\nIDN network together in an end-to-end fashion, our VCN network intimates\nprojection from the re-sampled vectors to the IDN-decoded image. As a result,\ngradients from IDN network to RSN network can be approximated by VCN network's\ngradient. Because dimension reduction can be further achieved by quantization\nin some dimensional space after image re-sampling within auto-encoder\narchitecture, we can well initialize our networks from pre-trained auto-encoder\nnetworks. Through extensive experiments and analysis, it is verified that the\nproposed method has more effectiveness and versatility than many\nstate-of-the-art approaches.","url_abs":"http://arxiv.org/abs/1806.08514v2","url_pdf":"http://arxiv.org/pdf/1806.08514v2.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":"virtual-codec-supervised-re-sampling-network","repo_url":"https://github.com/mdcnn/mdcnn.github.io","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"dimensionality-reduction","task_name":"Dimensionality Reduction"},{"task_slug":"image-compression","task_name":"Image Compression"},{"task_slug":"quantization","task_name":"Quantization"}],"methods":[{"method_slug":"discrete-cosine-transform","method_name":"Discrete Cosine Transform"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}