Research map / Label-efficient and robust learning / Recognition and long-tail
Multi-Scale Spatially-Asymmetric Recalibration for Image Classification
Yan Wang, Lingxi Xie, Siyuan Qiao, Ya Zhang, Wenjun Zhang, Alan L. Yuille
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 MS-SAR to ResNet-56 | ResNet-56 · CIFAR10 | error rate (%): 6.97 -> 6.04 | Improved |
| add MS-SAR to ResNet-56 | ResNet-56 · CIFAR100 | error rate (%): 29.07 -> 27.71 | Improved |
| add MS-SAR to DenseNet-100 | DenseNet-100 · CIFAR10 | error rate (%): 4.67 -> 4.16 | Improved |
| add MS-SAR to DenseNet-190 | DenseNet-190 · CIFAR10 | error rate (%): 3.46 -> 3.32 | Improved |
| add MS-SAR to ResNet-20 | ResNet-20 · CIFAR100 | error rate (%): 31.87 -> 31.09 | Improved |
| add scales 2 and 4 to scale 1 (multi-scale instead of single scale) (ablation) | MS-SAR on ResNet-56 · CIFAR10 | error rate (%): 6.80 -> 6.04 | Improved |
| add scale 4 to scale 2 (ablation) | MS-SAR on DenseNet-100 · CIFAR10 | error rate (%): 4.35 -> 4.33 (C10), 21.33 -> 21.23 (C100) | Hurt |
| use regional instead of sliding coordinate sets (ablation) | MS-SAR · CIFAR100 | error rate (%): regional outperforms sliding by ~0.2% | Improved |
| use multi-level instead of single-level recalibration (ablation) | MS-SAR on DenseNet-100 · CIFAR10 | error rate (%): 4.45 -> 4.06 | Improved |
| use multi-level instead of single-level recalibration (ablation) | MS-SAR on DenseNet-100 · CIFAR100 | error rate (%): 21.83 -> 21.13 | Improved |
| replace SE (single-scale symmetric) with MS-SAR | ResNet-18 · ILSVRC2012 val | top-1 error (%): 29.78 -> 29.43 | Improved |
| replace SE with MS-SAR | ResNeXt-50 · ILSVRC2012 val | top-5 error (%): 5.93 -> 5.78 | Improved |
Closest papers by meaning
- Asymmetric Convolution: An Efficient and Generalized Method to Fuse Feature Maps in Multiple Vision Tasks (TPAMI 2024 · similarity 0.84)
- Learning Asymmetric and Local Features in Multi-Dimensional Data Through Wavelets With Recursive Partitioning (TPAMI 2021 · similarity 0.84)
- AsCAN: Asymmetric Convolution-Attention Networks for Efficient Recognition and Generation (NeurIPS 2024 · similarity 0.82)
- Selective Disk Bispectrum: A Complete and Rotation Invariant Image Descriptor (NeurIPS 2026 · similarity 0.82)
- Multi-Scale VMamba: Hierarchy in Hierarchy Visual State Space Model (NeurIPS 2024 · similarity 0.82)
- Low-Rank Rescaled Vision Transformer Fine-Tuning: A Residual Design Approach (CVPR 2024 · similarity 0.81)
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.
