{"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/per-pixel-feedback-for-improving-semantic","title":"Per-Pixel Feedback for improving Semantic Segmentation","arxiv_id":"1712.02861","date":"2017-12-07","proceeding":null,"authors":["Aditya Ganeshan"],"abstract":"Semantic segmentation is the task of assigning a label to each pixel in the\nimage.In recent years, deep convolutional neural networks have been driving\nadvances in multiple tasks related to cognition. Although, DCNNs have resulted\nin unprecedented visual recognition performances, they offer little\ntransparency. To understand how DCNN based models work at the task of semantic\nsegmentation, we try to analyze the DCNN models in semantic segmentation. We\ntry to find the importance of global image information for labeling pixels.\n  Based on the experiments on discriminative regions, and modeling of\nfixations, we propose a set of new training loss functions for fine-tuning DCNN\nbased models. The proposed training regime has shown improvement in performance\nof DeepLab Large FOV(VGG-16) Segmentation model for PASCAL VOC 2012 dataset.\nHowever, further test remains to conclusively evaluate the benefits due to the\nproposed loss functions across models, and data-sets.\n  Submitted in part fulfillment of the requirements for the degree of\nIntegrated Masters of Science in Applied Mathematics.\n  Update: Further Experiment showed minimal benefits.\n  Code Available [here](https://github.com/BardOfCodes/Seg-Unravel).","url_abs":"http://arxiv.org/abs/1712.02861v1","url_pdf":"http://arxiv.org/pdf/1712.02861v1.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":"per-pixel-feedback-for-improving-semantic","repo_url":"https://github.com/BardOfCodes/Seg-Unravel","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"caffe2","reach":null}],"tasks":[{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[{"method_slug":"crf","method_name":"CRF"},{"method_slug":"dcnn","method_name":"DCNN"},{"method_slug":"deeplab","method_name":"DeepLab"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dilated-convolution","method_name":"Dilated Convolution"},{"method_slug":"feedforward-network","method_name":"Feedforward Network"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}