Papers › Discriminative Co-Saliency and Background Mining Transformer for Co-Salient Object Detection

Discriminative Co-Saliency and Background Mining Transformer for Co-Salient Object Detection

30 Apr 2023CVPR 2023 1arXiv:2305.00514archive 2025-07-28

Long Li, Junwei Han, Ni Zhang, Nian Liu, Salman Khan, Hisham Cholakkal, Rao Muhammad Anwer, Fahad Shahbaz Khan

Most previous co-salient object detection works mainly focus on extracting co-salient cues via mining the consistency relations across images while ignoring explicit exploration of background regions. In this paper, we propose a Discriminative co-saliency and background Mining Transformer framework (DMT) based on several economical multi-grained correlation modules to explicitly mine both co-saliency and background information and effectively model their discrimination. Specifically, we first propose a region-to-region correlation module for introducing inter-image relations to pixel-wise segmentation features while maintaining computational efficiency. Then, we use two types of pre-defined tokens to mine co-saliency and background information via our proposed contrast-induced pixel-to-token correlation and co-saliency token-to-token correlation modules. We also design a token-guided feature refinement module to enhance the discriminability of the segmentation features under the guidance of the learned tokens. We perform iterative mutual promotion for the segmentation feature extraction and token construction. Experimental results on three benchmark datasets demonstrate the effectiveness of our proposed method. The source code is available at: https://github.com/dragonlee258079/DMT.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

dragonlee258079/DMT officialmentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Co-Salient Object DetectionComputational EfficiencyObject DetectionSalient Object DetectionSegmentationobject-detection

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Co-Salient Object Detection CoCA DMT MAE 0.108 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoCA DMT Mean F-measure 0.590 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoCA DMT S-measure 0.725 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoCA DMT max E-measure 0.800 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoCA DMT max F-measure 0.619 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoCA DMT mean E-measure 0.753 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoSOD3k DMT MAE 0.063 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoSOD3k DMT S-measure 0.851 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoSOD3k DMT max E-measure 0.895 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoSOD3k DMT max F-measure 0.835 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoSOD3k DMT mean E-measure 0.881 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoSOD3k DMT mean F-measure 0.815 #2 of 10 Archive leaderboard report
Co-Salient Object Detection CoSal2015 DMT MAE 0.045 #1 of 10 Archive leaderboard report
Co-Salient Object Detection CoSal2015 DMT S-measure 0.897 #1 of 10 Archive leaderboard report
Co-Salient Object Detection CoSal2015 DMT max E-measure 0.936 #1 of 10 Archive leaderboard report
Co-Salient Object Detection CoSal2015 DMT max F-measure 0.905 #1 of 10 Archive leaderboard report
Co-Salient Object Detection CoSal2015 DMT mean E-measure 0.922 #1 of 10 Archive leaderboard report
Co-Salient Object Detection CoSal2015 DMT mean F-measure 0.883 #1 of 10 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections