Research map / Detection and segmentation / CNN features
HyperSeg: Patch-Wise Hypernetwork for Real-Time Semantic Segmentation
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 dynamic patch-wise convolutions instead of regular convolutions (ablation) | HyperSeg-S · CamVid test | mIoU: 77.3 -> 78.4 | Helped |
| Add dynamic patch-wise convolutions instead of regular convolutions (ablation) | HyperSeg-L · CamVid test | mIoU: 78.4 -> 79.1 | Helped |
| Replace EfficientNet-B1 with ResNet18 (ablation) | HyperSeg-S · CamVid test | mIoU: 78.4 -> 77.0 | Hurt |
| Replace EfficientNet-B1 with ResNet18 (ablation) | HyperSeg-L · CamVid test | mIoU: 79.1 -> 77.1 | Hurt |
| Replace EfficientNet-B1 with PSPNet18 (ablation) | HyperSeg-S · CamVid test | mIoU: 78.4 -> 76.6 | Hurt |
| Replace EfficientNet-B1 with PSPNet18 (ablation) | HyperSeg-L · CamVid test | mIoU: 79.1 -> 77.5 | Hurt |
| Replace EfficientNet-B1 with PSPNet50 (ablation) | HyperSeg-S · CamVid test | mIoU: 78.4 -> 77.1 | Hurt |
| Replace EfficientNet-B1 with PSPNet50 (ablation) | HyperSeg-L · CamVid test | mIoU: 79.1 -> 77.9 | Hurt |
| Increase the weight grid from 1×1 to 4×4 (ablation) | HyperSeg · PASCAL VOC 2012 val | mIoU: 77.56 -> 78.92 | Helped |
| Increase the weight grid from 4×4 to 8×8 (ablation) | HyperSeg · PASCAL VOC 2012 val | mIoU: 78.92 -> 80.23 | Helped |
| Increase the weight grid from 8×8 to 16×16 (ablation) | HyperSeg · PASCAL VOC 2012 val | mIoU: 80.23 -> 80.33 | Helped |
| Add positional encoding (ablation) | HyperSeg with a 16×16 weight grid · PASCAL VOC 2012 val | mIoU: 80.33 -> 80.61 | Helped |
Closest papers by meaning
- FasterSeg: Searching for Faster Real-time Semantic Segmentation (ICLR 2020 · similarity 0.92)
- HyperSeg: Hybrid Segmentation Assistant with Fine-grained Visual Perceiver (CVPR 2025 · similarity 0.88)
- PIDNet: A Real-Time Semantic Segmentation Network Inspired by PID Controllers (CVPR 2023 · similarity 0.88)
- RMAE-ProGRess: Advancing Semantic Segmentation in Unstructured Environments (CVPR 2026 · similarity 0.87)
- SCTNet: Single-Branch CNN with Transformer Semantic Information for Real-Time Segmentation (AAAI 2024 · similarity 0.87)
- Improving Semi-Supervised Semantic Segmentation with Sliced-Wasserstein Feature Alignment and Uniformity (CVPR 2025 · similarity 0.86)
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.