{"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/savoias-a-diverse-multi-category-visual","title":"SAVOIAS: A Diverse, Multi-Category Visual Complexity Dataset","arxiv_id":"1810.01771","date":"2018-10-03","proceeding":null,"authors":["Elham Saraee","Mona Jalal","Margrit Betke"],"abstract":"Visual complexity identifies the level of intricacy and details in an image\nor the level of difficulty to describe the image. It is an important concept in\na variety of areas such as cognitive psychology, computer vision and\nvisualization, and advertisement. Yet, efforts to create large, downloadable\nimage datasets with diverse content and unbiased groundtruthing are lacking. In\nthis work, we introduce Savoias, a visual complexity dataset that compromises\nof more than 1,400 images from seven image categories relevant to the above\nresearch areas, namely Scenes, Advertisements, Visualization and infographics,\nObjects, Interior design, Art, and Suprematism. The images in each category\nportray diverse characteristics including various low-level and high-level\nfeatures, objects, backgrounds, textures and patterns, text, and graphics. The\nground truth for Savoias is obtained by crowdsourcing more than 37,000 pairwise\ncomparisons of images using the forced-choice methodology and with more than\n1,600 contributors. The resulting relative scores are then converted to\nabsolute visual complexity scores using the Bradley-Terry method and matrix\ncompletion. When applying five state-of-the-art algorithms to analyze the\nvisual complexity of the images in the Savoias dataset, we found that the\nscores obtained from these baseline tools only correlate well with crowdsourced\nlabels for abstract patterns in the Suprematism category (Pearson correlation\nr=0.84). For the other categories, in particular, the objects and advertisement\ncategories, low correlation coefficients were revealed (r=0.3 and 0.56,\nrespectively). These findings suggest that (1) state-of-the-art approaches are\nmostly insufficient and (2) Savoias enables category-specific method\ndevelopment, which is likely to improve the impact of visual complexity\nanalysis on specific application areas, including computer vision.","url_abs":"http://arxiv.org/abs/1810.01771v1","url_pdf":"http://arxiv.org/pdf/1810.01771v1.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":"savoias-a-diverse-multi-category-visual","repo_url":"https://github.com/esaraee/Savoias-Dataset","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"matrix-completion","task_name":"Matrix Completion"}],"methods":[],"datasets_introduced":[{"slug":"savoias","name":"SAVOIAS","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}