{"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/learning-to-exploit-the-prior-network","title":"Learning to Exploit the Prior Network Knowledge for Weakly-Supervised Semantic Segmentation","arxiv_id":"1804.04882","date":"2018-04-13","proceeding":null,"authors":["Carolina Redondo-Cabrera","Marcos Baptista-Ríos","Roberto J. López-Sastre"],"abstract":"Training a Convolutional Neural Network (CNN) for semantic segmentation\ntypically requires to collect a large amount of accurate pixel-level\nannotations, a hard and expensive task. In contrast, simple image tags are\neasier to gather. With this paper we introduce a novel weakly-supervised\nsemantic segmentation model able to learn from image labels, and just image\nlabels. Our model uses the prior knowledge of a network trained for image\nrecognition, employing these image annotations as an attention mechanism to\nidentify semantic regions in the images. We then present a methodology that\nbuilds accurate class-specific segmentation masks from these regions, where\nneither external objectness nor saliency algorithms are required. We describe\nhow to incorporate this mask generation strategy into a fully end-to-end\ntrainable process where the network jointly learns to classify and segment\nimages. Our experiments on PASCAL VOC 2012 dataset show that exploiting these\ngenerated class-specific masks in conjunction with our novel end-to-end\nlearning process outperforms several recent weakly-supervised semantic\nsegmentation methods that use image tags only, and even some models that\nleverage additional supervision or training data.","url_abs":"http://arxiv.org/abs/1804.04882v2","url_pdf":"http://arxiv.org/pdf/1804.04882v2.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":"learning-to-exploit-the-prior-network","repo_url":"https://github.com/gramuah/weakly-supervised-segmentation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"weakly-supervised-semantic-segmentation-1","task_name":"Weakly supervised Semantic Segmentation"},{"task_slug":"weakly-supervised-semantic-segmentation","task_name":"Weakly-Supervised Semantic Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}