Research map / Detection and segmentation / CNN features
Learning Uncertain Convolutional Features for Accurate Saliency Detection
Pingping Zhang, Dong Wang, Huchuan Lu, Hongyu Wang, Baocai Yin
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace plain dropout with R-Dropout (ablation) | UCF (regular deconvolution) · ECSSD | F-measure: 0.8166 -> 0.8356 | Improved |
| replace plain dropout with R-Dropout (ablation) | UCF (regular deconvolution) · ECSSD | MAE: 0.0811 -> 0.0781 | Improved |
| replace plain dropout with R-Dropout (ablation) | UCF (regular deconvolution) · DUT-OMRON | F-measure: 0.6128 -> 0.6168 | Improved |
| replace plain dropout with R-Dropout (ablation) | UCF (regular deconvolution) · PASCAL-S | F-measure: 0.6172 -> 0.6845 | Improved |
| replace regular deconvolution with interpolation upsampling (ablation) | UCF with R-Dropout · ECSSD | F-measure: 0.8356 -> 0.8426 | Improved |
| replace regular deconvolution with interpolation upsampling (ablation) | UCF with R-Dropout · PASCAL-S | F-measure: 0.6845 -> 0.7201 | Improved |
| use hybrid upsampling instead of interpolation only (ablation) | UCF with R-Dropout · ECSSD | F-measure: 0.8426 -> 0.8517 | Improved |
| use hybrid upsampling instead of deconvolution only (ablation) | UCF with R-Dropout · ECSSD | F-measure: 0.8356 -> 0.8517 | Improved |
| use hybrid upsampling instead of interpolation only (ablation) | UCF with R-Dropout · SED1 | F-measure: 0.8665 -> 0.8647 | No clear effect |
| add R-Dropout to interpolation-only upsampling model (ablation) | UCF without dropout · ECSSD | F-measure: 0.7857 -> 0.8426 | Improved |
| add R-Dropout to deconvolution-only upsampling model (ablation) | UCF without dropout · ECSSD | F-measure: 0.6942 -> 0.8356 | Improved |
| use hybrid upsampling instead of regular deconvolution (ablation) | UCF · PASCAL VOC 2012 segmentation | mean IoU: 65.173 -> 67.45 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- RFCN — R-FCN: Object Detection via Region-based Fully Convolutional Networks (NeurIPS 2016) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Non-Local Deep Features for Salient Object Detection (CVPR 2017 · similarity 0.89)
- Recurrently Aggregating Deep Features for Salient Object Detection (AAAI 2018 · similarity 0.88)
- Deep Contrast Learning for Salient Object Detection (CVPR 2016 · similarity 0.88)
- Salient Object Detection with Recurrent Fully Convolutional Networks (TPAMI 2018 · similarity 0.88)
- Multi-Source Uncertainty Mining for Deep Unsupervised Saliency Detection (CVPR 2022 · similarity 0.88)
- UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders (CVPR 2020 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
