Research map / Detection and segmentation / Weakly supervised detection and segmentation
Deep Dual Learning for Semantic Image Segmentation
Ping Luo, Guangrun Wang, Liang Lin, Xiaogang Wang
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 |
|---|---|---|---|
| add dual-learning inference refinement of labelmaps and tags during training (ttr=0 to ttr=5, tts=0) (ablation) | DIS · VOC12 test | mIoU: 80.0 -> 84.6 | Improved |
| increase training inference iterations from ttr=5 to ttr=30 with tts=0 (ablation) | DIS · VOC12 test | mIoU: 84.6 -> 86.3 | Improved |
| increase test inference iterations from tts=0 to tts=30 at ttr=5 (ablation) | DIS · VOC12 test | mIoU: 84.6 -> 86.2 | Improved |
| increase test inference iterations tts=0 to tts=30 at ttr=10 (ablation) | DIS · VOC12 test | mIoU: 85.5 -> 86.4 | Improved |
| increase test inference iterations tts=0 to tts=30 at ttr=30 (ablation) | DIS · VOC12 test | mIoU: 86.3 -> 86.8 | Improved |
| increase ttr from 5 to 30 at tts=10 (ablation) | DIS · VOC12 test | mIoU: 85.4 -> 86.5 | Improved |
| semi-supervised scheme without refinement (ttr=0) vs ResNet101 baseline (ablation) | ResNet101 baseline · VOC12 test | mIoU: outperforms the baseline by 5.8% | Improved |
| replace WSSL+IDW with DIS using the same weak data | WSSL+IDW · VOC12 test | mIoU: 81.9 -> 86.8 | Improved |
| add IDW weak tags to WSSL | WSSL (semi) · VOC12 test | mIoU: 69.0 -> 81.9 | Improved |
| increase training inference iterations from ttr=10 to ttr=30 at tts=0 (ablation) | DIS · IDW test | mIoU: 56.6 -> 57.5 | Improved |
| increase test iterations tts=0 to tts=30 at ttr=30 (ablation) | DIS · IDW test | mIoU: 57.5 -> 59.8 | Improved |
| DIS ttr=30, tts=30 vs WSSL+IDW | WSSL+IDW · IDW test | mIoU: 52.0 -> 59.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- SegNet — SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation (TPAMI 2017) · better on 1 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
- DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs (TPAMI 2017 · similarity 0.88)
- Combining Bottom-Up, Top-Down, and Smoothness Cues for Weakly Supervised Image Segmentation (CVPR 2017 · similarity 0.88)
- On the Importance of Label Quality for Semantic Segmentation (CVPR 2018 · similarity 0.88)
- Semantic Image Segmentation With Task-Specific Edge Detection Using CNNs and a Discriminatively Trained Domain Transform (CVPR 2016 · similarity 0.87)
- Learning Pixel-Level Semantic Affinity With Image-Level Supervision for Weakly Supervised Semantic Segmentation (CVPR 2018 · similarity 0.87)
- Improving Semi-Supervised and Domain-Adaptive Semantic Segmentation with Self-Supervised Depth Estimation (IJCV 2023 · 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.
