{"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/macro-micro-adversarial-network-for-human","title":"Macro-Micro Adversarial Network for Human Parsing","arxiv_id":"1807.08260","date":"2018-07-22","proceeding":"ECCV 2018 9","authors":["Yawei Luo","Zhedong Zheng","Liang Zheng","Tao Guan","Junqing Yu","Yi Yang"],"abstract":"In human parsing, the pixel-wise classification loss has drawbacks in its\nlow-level local inconsistency and high-level semantic inconsistency. The\nintroduction of the adversarial network tackles the two problems using a single\ndiscriminator. However, the two types of parsing inconsistency are generated by\ndistinct mechanisms, so it is difficult for a single discriminator to solve\nthem both. To address the two kinds of inconsistencies, this paper proposes the\nMacro-Micro Adversarial Net (MMAN). It has two discriminators. One\ndiscriminator, Macro D, acts on the low-resolution label map and penalizes\nsemantic inconsistency, e.g., misplaced body parts. The other discriminator,\nMicro D, focuses on multiple patches of the high-resolution label map to\naddress the local inconsistency, e.g., blur and hole. Compared with traditional\nadversarial networks, MMAN not only enforces local and semantic consistency\nexplicitly, but also avoids the poor convergence problem of adversarial\nnetworks when handling high resolution images. In our experiment, we validate\nthat the two discriminators are complementary to each other in improving the\nhuman parsing accuracy. The proposed framework is capable of producing\ncompetitive parsing performance compared with the state-of-the-art methods,\ni.e., mIoU=46.81% and 59.91% on LIP and PASCAL-Person-Part, respectively. On a\nrelatively small dataset PPSS, our pre-trained model demonstrates impressive\ngeneralization ability. The code is publicly available at\nhttps://github.com/RoyalVane/MMAN.","url_abs":"http://arxiv.org/abs/1807.08260v2","url_pdf":"http://arxiv.org/pdf/1807.08260v2.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":"macro-micro-adversarial-network-for-human","repo_url":"https://github.com/RoyalVane/MMAN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"human-parsing","task_name":"Human Parsing"},{"task_slug":"human-part-segmentation","task_name":"Human Part Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/semantic-segmentation-on-lip-val","task":"Semantic Segmentation","dataset":"LIP val","model":"MMAN (ResNet-101)","rank_in_archive_order":12,"of":13,"metrics":{"mIoU":"46.81%"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.08260","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}