{"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/the-devil-is-in-the-decoder-classification","title":"The Devil is in the Decoder: Classification, Regression and GANs","arxiv_id":"1707.05847","date":"2017-07-18","proceeding":null,"authors":["Zbigniew Wojna","Vittorio Ferrari","Sergio Guadarrama","Nathan Silberman","Liang-Chieh Chen","Alireza Fathi","Jasper Uijlings"],"abstract":"Many machine vision applications, such as semantic segmentation and depth\nprediction, require predictions for every pixel of the input image. Models for\nsuch problems usually consist of encoders which decrease spatial resolution\nwhile learning a high-dimensional representation, followed by decoders who\nrecover the original input resolution and result in low-dimensional\npredictions. While encoders have been studied rigorously, relatively few\nstudies address the decoder side. This paper presents an extensive comparison\nof a variety of decoders for a variety of pixel-wise tasks ranging from\nclassification, regression to synthesis. Our contributions are: (1) Decoders\nmatter: we observe significant variance in results between different types of\ndecoders on various problems. (2) We introduce new residual-like connections\nfor decoders. (3) We introduce a novel decoder: bilinear additive upsampling.\n(4) We explore prediction artifacts.","url_abs":"http://arxiv.org/abs/1707.05847v3","url_pdf":"http://arxiv.org/pdf/1707.05847v3.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":"the-devil-is-in-the-decoder-classification","repo_url":"https://github.com/bayraktarbaris/SNGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"boundary-detection","task_name":"Boundary Detection"},{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"depth-estimation","task_name":"Depth Estimation"},{"task_slug":"depth-prediction","task_name":"Depth Prediction"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1707.05847","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}