{"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/box-driven-class-wise-region-masking-and","title":"Box-driven Class-wise Region Masking and Filling Rate Guided Loss for Weakly Supervised Semantic Segmentation","arxiv_id":"1904.11693","date":"2019-04-26","proceeding":"CVPR 2019 6","authors":["Chunfeng Song","Yan Huang","Wanli Ouyang","Liang Wang"],"abstract":"Semantic segmentation has achieved huge progress via adopting deep Fully\nConvolutional Networks (FCN). However, the performance of FCN based models\nseverely rely on the amounts of pixel-level annotations which are expensive and\ntime-consuming. To address this problem, it is a good choice to learn to\nsegment with weak supervision from bounding boxes. How to make full use of the\nclass-level and region-level supervisions from bounding boxes is the critical\nchallenge for the weakly supervised learning task. In this paper, we first\nintroduce a box-driven class-wise masking model (BCM) to remove irrelevant\nregions of each class. Moreover, based on the pixel-level segment proposal\ngenerated from the bounding box supervision, we could calculate the mean\nfilling rates of each class to serve as an important prior cue, then we propose\na filling rate guided adaptive loss (FR-Loss) to help the model ignore the\nwrongly labeled pixels in proposals. Unlike previous methods directly training\nmodels with the fixed individual segment proposals, our method can adjust the\nmodel learning with global statistical information. Thus it can help reduce the\nnegative impacts from wrongly labeled proposals. We evaluate the proposed\nmethod on the challenging PASCAL VOC 2012 benchmark and compare with other\nmethods. Extensive experimental results show that the proposed method is\neffective and achieves the state-of-the-art results.","url_abs":"http://arxiv.org/abs/1904.11693v1","url_pdf":"http://arxiv.org/pdf/1904.11693v1.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":"box-driven-class-wise-region-masking-and","repo_url":"https://github.com/developfeng/BCM","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"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"},{"task_slug":"weakly-supervised-learning","task_name":"Weakly-supervised Learning"}],"methods":[{"method_slug":"adaptive-loss","method_name":"Adaptive Loss"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"fcn","method_name":"FCN"},{"method_slug":"max-pooling","method_name":"Max Pooling"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1904.11693","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}