{"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/every-filter-extracts-a-specific-texture-in","title":"Every Filter Extracts A Specific Texture In Convolutional Neural Networks","arxiv_id":"1608.04170","date":"2016-08-15","proceeding":null,"authors":["Zhiqiang Xia","Ce Zhu","Zhengtao Wang","Qi Guo","Yipeng Liu"],"abstract":"Many works have concentrated on visualizing and understanding the inner\nmechanism of convolutional neural networks (CNNs) by generating images that\nactivate some specific neurons, which is called deep visualization. However, it\nis still unclear what the filters extract from images intuitively. In this\npaper, we propose a modified code inversion algorithm, called feature map\ninversion, to understand the function of filter of interest in CNNs. We reveal\nthat every filter extracts a specific texture. The texture from higher layer\ncontains more colours and more intricate structures. We also demonstrate that\nstyle of images could be a combination of these texture primitives. Two methods\nare proposed to reallocate energy distribution of feature maps randomly and\npurposefully. Then, we inverse the modified code and generate images of diverse\nstyles. With these results, we provide an explanation about why Gram matrix of\nfeature maps \\cite{Gatys_2016_CVPR} could represent image style.","url_abs":"http://arxiv.org/abs/1608.04170v2","url_pdf":"http://arxiv.org/pdf/1608.04170v2.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":"every-filter-extracts-a-specific-texture-in","repo_url":"https://github.com/xzqjack/FeatureMapInversion","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"mxnet","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}