{"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/deep-edge-aware-saliency-detection","title":"Deep Edge-Aware Saliency Detection","arxiv_id":"1708.04366","date":"2017-08-15","proceeding":null,"authors":["Jing Zhang","Yuchao Dai","Fatih Porikli","Mingyi He"],"abstract":"There has been profound progress in visual saliency thanks to the deep\nlearning architectures, however, there still exist three major challenges that\nhinder the detection performance for scenes with complex compositions, multiple\nsalient objects, and salient objects of diverse scales. In particular, output\nmaps of the existing methods remain low in spatial resolution causing blurred\nedges due to the stride and pooling operations, networks often neglect\ndescriptive statistical and handcrafted priors that have potential to\ncomplement saliency detection results, and deep features at different layers\nstay mainly desolate waiting to be effectively fused to handle multi-scale\nsalient objects. In this paper, we tackle these issues by a new fully\nconvolutional neural network that jointly learns salient edges and saliency\nlabels in an end-to-end fashion. Our framework first employs convolutional\nlayers that reformulate the detection task as a dense labeling problem, then\nintegrates handcrafted saliency features in a hierarchical manner into lower\nand higher levels of the deep network to leverage available information for\nmulti-scale response, and finally refines the saliency map through dilated\nconvolutions by imposing context. In this way, the salient edge priors are\nefficiently incorporated and the output resolution is significantly improved\nwhile keeping the memory requirements low, leading to cleaner and sharper\nobject boundaries. Extensive experimental analyses on ten benchmarks\ndemonstrate that our framework achieves consistently superior performance and\nattains robustness for complex scenes in comparison to the very recent\nstate-of-the-art approaches.","url_abs":"http://arxiv.org/abs/1708.04366v1","url_pdf":"http://arxiv.org/pdf/1708.04366v1.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":"deep-edge-aware-saliency-detection","repo_url":"https://github.com/MindSpore-scientific/code-12/tree/main/Deep-Edge-Aware-Interactive","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"descriptive","task_name":"Descriptive"},{"task_slug":"saliency-detection","task_name":"Saliency Detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}