{"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/segnet-a-deep-convolutional-encoder-decoder-1","title":"SegNet: A Deep Convolutional Encoder-Decoder Architecture for Robust Semantic Pixel-Wise Labelling","arxiv_id":"1505.07293","date":"2015-05-27","proceeding":null,"authors":["Vijay Badrinarayanan","Ankur Handa","Roberto Cipolla"],"abstract":"We propose a novel deep architecture, SegNet, for semantic pixel wise image\nlabelling. SegNet has several attractive properties; (i) it only requires\nforward evaluation of a fully learnt function to obtain smooth label\npredictions, (ii) with increasing depth, a larger context is considered for\npixel labelling which improves accuracy, and (iii) it is easy to visualise the\neffect of feature activation(s) in the pixel label space at any depth. SegNet\nis composed of a stack of encoders followed by a corresponding decoder stack\nwhich feeds into a soft-max classification layer. The decoders help map low\nresolution feature maps at the output of the encoder stack to full input image\nsize feature maps. This addresses an important drawback of recent deep learning\napproaches which have adopted networks designed for object categorization for\npixel wise labelling. These methods lack a mechanism to map deep layer feature\nmaps to input dimensions. They resort to ad hoc methods to upsample features,\ne.g. by replication. This results in noisy predictions and also restricts the\nnumber of pooling layers in order to avoid too much upsampling and thus reduces\nspatial context. SegNet overcomes these problems by learning to map encoder\noutputs to image pixel labels. We test the performance of SegNet on outdoor RGB\nscenes from CamVid, KITTI and indoor scenes from the NYU dataset. Our results\nshow that SegNet achieves state-of-the-art performance even without use of\nadditional cues such as depth, video frames or post-processing with CRF models.","url_abs":"http://arxiv.org/abs/1505.07293v1","url_pdf":"http://arxiv.org/pdf/1505.07293v1.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":"segnet-a-deep-convolutional-encoder-decoder-1","repo_url":"https://github.com/IRVLab/SUIM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"segnet-a-deep-convolutional-encoder-decoder-1","repo_url":"https://github.com/IRVLab/SUIM-Net","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"segnet-a-deep-convolutional-encoder-decoder-1","repo_url":"https://github.com/keillernogueira/FDSI","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"segnet-a-deep-convolutional-encoder-decoder-1","repo_url":"https://github.com/xahidbuffon/SUIM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"segnet-a-deep-convolutional-encoder-decoder-1","repo_url":"https://github.com/xahidbuffon/SVAM-Net","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"object-categorization","task_name":"Object Categorization"}],"methods":[{"method_slug":"crf","method_name":"CRF"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1505.07293","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}