{"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/hierarchical-cellular-automata-for-visual","title":"Hierarchical Cellular Automata for Visual Saliency","arxiv_id":"1705.09425","date":"2017-05-26","proceeding":null,"authors":["Yao Qin","Mengyang Feng","Huchuan Lu","Garrison W. Cottrell"],"abstract":"Saliency detection, finding the most important parts of an image, has become\nincreasingly popular in computer vision. In this paper, we introduce\nHierarchical Cellular Automata (HCA) -- a temporally evolving model to\nintelligently detect salient objects. HCA consists of two main components:\nSingle-layer Cellular Automata (SCA) and Cuboid Cellular Automata (CCA). As an\nunsupervised propagation mechanism, Single-layer Cellular Automata can exploit\nthe intrinsic relevance of similar regions through interactions with neighbors.\nLow-level image features as well as high-level semantic information extracted\nfrom deep neural networks are incorporated into the SCA to measure the\ncorrelation between different image patches. With these hierarchical deep\nfeatures, an impact factor matrix and a coherence matrix are constructed to\nbalance the influences on each cell's next state. The saliency values of all\ncells are iteratively updated according to a well-defined update rule.\nFurthermore, we propose CCA to integrate multiple saliency maps generated by\nSCA at different scales in a Bayesian framework. Therefore, single-layer\npropagation and multi-layer integration are jointly modeled in our unified HCA.\nSurprisingly, we find that the SCA can improve all existing methods that we\napplied it to, resulting in a similar precision level regardless of the\noriginal results. The CCA can act as an efficient pixel-wise aggregation\nalgorithm that can integrate state-of-the-art methods, resulting in even better\nresults. Extensive experiments on four challenging datasets demonstrate that\nthe proposed algorithm outperforms state-of-the-art conventional methods and is\ncompetitive with deep learning based approaches.","url_abs":"http://arxiv.org/abs/1705.09425v1","url_pdf":"http://arxiv.org/pdf/1705.09425v1.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":"hierarchical-cellular-automata-for-visual","repo_url":"https://github.com/ArcherFMY/HCA_saliency_codes","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"saliency-detection","task_name":"Saliency Detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}