{"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/exploiting-high-level-semantics-for-no","title":"Exploiting High-Level Semantics for No-Reference Image Quality Assessment of Realistic Blur Images","arxiv_id":"1810.08169","date":"2018-10-18","proceeding":null,"authors":["Dingquan Li","Tingting Jiang","Ming Jiang"],"abstract":"To guarantee a satisfying Quality of Experience (QoE) for consumers, it is\nrequired to measure image quality efficiently and reliably. The neglect of the\nhigh-level semantic information may result in predicting a clear blue sky as\nbad quality, which is inconsistent with human perception. Therefore, in this\npaper, we tackle this problem by exploiting the high-level semantics and\npropose a novel no-reference image quality assessment method for realistic blur\nimages. Firstly, the whole image is divided into multiple overlapping patches.\nSecondly, each patch is represented by the high-level feature extracted from\nthe pre-trained deep convolutional neural network model. Thirdly, three\ndifferent kinds of statistical structures are adopted to aggregate the\ninformation from different patches, which mainly contain some common statistics\n(i.e., the mean\\&standard deviation, quantiles and moments). Finally, the\naggregated features are fed into a linear regression model to predict the image\nquality. Experiments show that, compared with low-level features, high-level\nfeatures indeed play a more critical role in resolving the aforementioned\nchallenging problem for quality estimation. Besides, the proposed method\nsignificantly outperforms the state-of-the-art methods on two realistic blur\nimage databases and achieves comparable performance on two synthetic blur image\ndatabases.","url_abs":"http://arxiv.org/abs/1810.08169v1","url_pdf":"http://arxiv.org/pdf/1810.08169v1.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":"exploiting-high-level-semantics-for-no","repo_url":"https://github.com/lidq92/SFA","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"image-quality-assessment","task_name":"Image Quality Assessment"},{"task_slug":"no-reference-image-quality-assessment","task_name":"No-Reference Image Quality Assessment"}],"methods":[{"method_slug":"linear-regression","method_name":"Linear Regression"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}