{"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/picanet-pixel-wise-contextual-attention","title":"PiCANet: Pixel-wise Contextual Attention Learning for Accurate Saliency Detection","arxiv_id":"1812.06314","date":"2018-12-15","proceeding":null,"authors":["Nian Liu","Junwei Han","Ming-Hsuan Yang"],"abstract":"In saliency detection, every pixel needs contextual information to make\nsaliency prediction. Previous models usually incorporate contexts holistically.\nHowever, for each pixel, usually only part of its context region is useful and\ncontributes to its prediction, while some other part may serve as noises and\ndistractions. In this paper, we propose a novel pixel-wise contextual attention\nnetwork, \\ie PiCANet, to learn to selectively attend to informative context\nlocations at each pixel. Specifically, PiCANet generates an attention map over\nthe context region of each pixel, where each attention weight corresponds to\nthe relevance of a context location w.r.t the referred pixel. Then, attentive\ncontextual features can be constructed via selectively incorporating the\nfeatures of useful context locations with the learned attention. We propose\nthree specific formulations of the PiCANet via embedding the pixel-wise\ncontextual attention mechanism into the pooling and convolution operations with\nattending to global or local contexts. All the three models are fully\ndifferentiable and can be integrated with CNNs with joint training. We\nintroduce the proposed PiCANets into a U-Net architecture for salient object\ndetection. Experimental results indicate that the proposed PiCANets can\nsignificantly improve the saliency detection performance. The generated global\nand local attention can learn to incorporate global contrast and smoothness,\nrespectively, which help localize salient objects more accurately and highlight\nthem more uniformly. Consequently, our saliency model performs favorably\nagainst other state-of-the-art methods. Moreover, we also validate that\nPiCANets can also improve semantic segmentation and object detection\nperformances, which further demonstrates their effectiveness and generalization\nability.","url_abs":"http://arxiv.org/abs/1812.06314v1","url_pdf":"http://arxiv.org/pdf/1812.06314v1.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":"picanet-pixel-wise-contextual-attention","repo_url":"https://github.com/nian-liu/PiCANet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null},{"paper_slug":"picanet-pixel-wise-contextual-attention","repo_url":"https://github.com/wangxiao5791509/cmSalGAN_PyTorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"salient-object-detection","task_name":"RGB Salient Object Detection"},{"task_slug":"saliency-detection","task_name":"Saliency Detection"},{"task_slug":"saliency-prediction","task_name":"Saliency Prediction"},{"task_slug":"salient-object-detection-1","task_name":"Salient Object Detection"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"concatenated-skip-connection","method_name":"Concatenated Skip Connection"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"u-net","method_name":"U-Net"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1812.06314","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}