Research map / Detection and segmentation / CNN features
SeeThroughNet: Resurrection of Auxiliary Loss by Preserving Class Probability Information
Dasol Han, Jaewook Yoo, Dokwan Oh
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 CPP pooling to auxiliary heads | OCRNet · Cityscapes val | mIoU (%): 81.17 -> 81.79 | Improved |
| Add CPP pooling to auxiliary heads | DeepLabV3+ · Pascal VOC 2012 val | mIoU (%): 78.35 -> 79.89 | Improved |
| Add CPP pooling to auxiliary heads | ANN · Pascal Context val | mIoU (%): 44.42 -> 47.19 | Improved |
| Add CPP pooling to auxiliary heads | OCRNet · NYU-Depth-v2 val | mIoU (%): 58.09 -> 58.59 | Improved |
| Add CPP pooling (ablation) | SeeThroughNet · Cityscapes val | IoU (%): 84.06 -> 84.69 | Improved |
| Add multi-scale attention masks (ablation) | SeeThroughNet · Cityscapes val | IoU (%): 84.06 -> 84.37 | Improved |
| Combine CPP pooling with multi-scale attention masks (ablation) | SeeThroughNet · Cityscapes val | IoU (%): 84.06 -> 85.29 | Improved |
| Use CPP pooling with a multi-scale attention-coupled decoder | SeeThroughNet · Cityscapes test | mIoU (%): 85.61 -> 86.12 | Improved |
Closest papers by meaning
- Loss Max-Pooling for Semantic Image Segmentation (CVPR 2017 · similarity 0.87)
- InverseForm: A Loss Function for Structured Boundary-Aware Segmentation (CVPR 2021 · similarity 0.85)
- Building Detail-Sensitive Semantic Segmentation Networks With Polynomial Pooling (CVPR 2019 · similarity 0.85)
- PIDNet: A Real-Time Semantic Segmentation Network Inspired by PID Controllers (CVPR 2023 · similarity 0.85)
- Consistent-Separable Feature Representation for Semantic Segmentation (AAAI 2021 · similarity 0.85)
- Complex Wavelet Mutual Information Loss: A Multi-Scale Loss Function for Semantic Segmentation (ICML 2025 · similarity 0.83)
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.
