{"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/cascaded-partial-decoder-for-fast-and","title":"Cascaded Partial Decoder for Fast and Accurate Salient Object Detection","arxiv_id":"1904.08739","date":"2019-04-18","proceeding":"CVPR 2019 6","authors":["Zhe Wu","Li Su","Qingming Huang"],"abstract":"Existing state-of-the-art salient object detection networks rely on\naggregating multi-level features of pre-trained convolutional neural networks\n(CNNs). Compared to high-level features, low-level features contribute less to\nperformance but cost more computations because of their larger spatial\nresolutions. In this paper, we propose a novel Cascaded Partial Decoder (CPD)\nframework for fast and accurate salient object detection. On the one hand, the\nframework constructs partial decoder which discards larger resolution features\nof shallower layers for acceleration. On the other hand, we observe that\nintegrating features of deeper layers obtain relatively precise saliency map.\nTherefore we directly utilize generated saliency map to refine the features of\nbackbone network. This strategy efficiently suppresses distractors in the\nfeatures and significantly improves their representation ability. Experiments\nconducted on five benchmark datasets exhibit that the proposed model not only\nachieves state-of-the-art performance but also runs much faster than existing\nmodels. Besides, the proposed framework is further applied to improve existing\nmulti-level feature aggregation models and significantly improve their\nefficiency and accuracy.","url_abs":"http://arxiv.org/abs/1904.08739v1","url_pdf":"http://arxiv.org/pdf/1904.08739v1.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":"cascaded-partial-decoder-for-fast-and","repo_url":"https://github.com/wuzhe71/CPD","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"camouflaged-object-segmentation","task_name":"Camouflaged Object Segmentation"},{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"salient-object-detection","task_name":"RGB Salient Object Detection"},{"task_slug":"salient-object-detection-1","task_name":"Salient Object Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/camouflaged-object-segmentation-on-cod","task":"Camouflaged Object Segmentation","dataset":"COD","model":"CPD","rank_in_archive_order":9,"of":12,"metrics":{"MAE":"0.059","S-Measure":"0.747","Weighted F-Measure":"0.508"},"uses_additional_data":false},{"leaderboard":"/sota/camouflaged-object-segmentation-on-pcod-1200","task":"Camouflaged Object Segmentation","dataset":"PCOD_1200","model":"CPD","rank_in_archive_order":13,"of":16,"metrics":{"S-Measure":"0.855"},"uses_additional_data":false},{"leaderboard":"/sota/salient-object-detection-on-dut-omron","task":"RGB Salient Object Detection","dataset":"DUT-OMRON","model":"CPD-R (ResNet50)","rank_in_archive_order":14,"of":18,"metrics":{"F-measure":"0.747","MAE":"0.056"},"uses_additional_data":false},{"leaderboard":"/sota/salient-object-detection-on-ecssd","task":"RGB Salient Object Detection","dataset":"ECSSD","model":"CPD-R (ResNet50)","rank_in_archive_order":10,"of":14,"metrics":{"F-measure":"0.917","MAE":"0.037"},"uses_additional_data":false},{"leaderboard":"/sota/salient-object-detection-on-hku-is","task":"RGB Salient Object Detection","dataset":"HKU-IS","model":"CPD-R (ResNet50)","rank_in_archive_order":10,"of":14,"metrics":{"F-measure":"0.891","MAE":"0.034"},"uses_additional_data":false},{"leaderboard":"/sota/salient-object-detection-on-istd","task":"RGB Salient Object Detection","dataset":"ISTD","model":"CPD","rank_in_archive_order":1,"of":7,"metrics":{"Balanced Error Rate":"6.76"},"uses_additional_data":false},{"leaderboard":"/sota/salient-object-detection-on-pascal-s","task":"RGB Salient Object Detection","dataset":"PASCAL-S","model":"CPD-R (ResNet50)","rank_in_archive_order":9,"of":13,"metrics":{"F-measure":"0.824","MAE":"0.072"},"uses_additional_data":false},{"leaderboard":"/sota/salient-object-detection-on-sbu","task":"RGB Salient Object Detection","dataset":"SBU / SBU-Refine","model":"CPD","rank_in_archive_order":1,"of":7,"metrics":{"Balanced Error Rate":"4.19"},"uses_additional_data":false},{"leaderboard":"/sota/salient-object-detection-on-ucf","task":"RGB Salient Object Detection","dataset":"UCF","model":"CPD","rank_in_archive_order":1,"of":7,"metrics":{"Balanced Error Rate":"7.21"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1904.08739","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}