Research map / Detection and segmentation / CNN features
Deeply Supervised Salient Object Detection With Short Connections
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Add short connections using Pattern 1 (ablation) | Enhanced HED · PASCALS | F-measure: no performance gain | No clear effect |
| Use short connection Pattern 2 (ablation) | Enhanced HED · PASCALS | F-measure: +0.8 points | Helped |
| Use short connection Pattern 3 (ablation) | Enhanced HED · PASCALS | F-measure: +0.6 points | Helped |
| Add a pool5 side output and two additional convolutional layers per side output (ablation) | HED · Salient object detection | F-measure: about $3\%$ improvement | Helped |
| Increase the number of channels in each side output (ablation) | DSS · PASCALS | F-measure | No clear effect |
| Remove one convolutional layer from each side output (ablation) | DSS · PASCALS | F-measure: decreases the performance by 1.5 points | Hurt |
| Reduce the kernel size in deeper side outputs (ablation) | DSS · PASCALS | F-measure | Hurt |
| Use the hole algorithm to densify deep side-output maps (ablation) | DSS · Validation set | F-measure: decreased by nearly 1% | Hurt |
| Horizontally flip all training images (ablation) | DSS · Salient object detection test sets | F-measure: improves the performance by more than 0.5% | Helped |
| Crop inputs to 321×321 instead of training on full-resolution images (ablation) | DSS · Salient object detection test sets | F-measure: decrease our performance by more than 0.5 points | Hurt |
| Replace VGGNet with ResNet-101 (ablation) | DSS · Five salient object detection datasets | F-measure: one-point improvement on each dataset in terms of F-measure score on average | Helped |
| Add a modulating factor to the CRF unary term (ablation) | DSS with CRF · Salient object detection test sets | MAE: reduce the MAE scores by around 0.3 points | Helped |
Closest papers by meaning
- Salient Object Detection With Pyramid Attention and Salient Edges (CVPR 2019 · similarity 0.90)
- A Mutual Learning Method for Salient Object Detection With Intertwined Multi-Supervision (CVPR 2019 · similarity 0.89)
- Deep Embedding Features for Salient Object Detection (AAAI 2019 · similarity 0.89)
- Weakly Supervised Salient Object Detection Using Image Labels (AAAI 2018 · similarity 0.89)
- Recurrently Aggregating Deep Features for Salient Object Detection (AAAI 2018 · similarity 0.88)
- An Iterative and Cooperative Top-Down and Bottom-Up Inference Network for Salient Object Detection (CVPR 2019 · similarity 0.88)
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.