Research map / Image generation / Image restoration
SinBasis Networks: Matrix-Equivalent Feature Extraction for Wave-Like Optical Spectrograms
Yuzhou Zhu, Zheng Zhang, Ruyi Zhang, Liang Zhou
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 |
|---|---|---|---|
| apply sin() to weight matrices | CNN · Synthetic attosecond spectrograms | MSE: 0.0120 -> 0.0080 | Improved |
| apply sin() to patch-embedding weights | ViT · Synthetic attosecond spectrograms | MSE: 0.0100 -> 0.0070 | Improved |
| apply sin() to capsule vote matrices | Capsule · Synthetic attosecond spectrograms | MSE: 0.0110 -> 0.0060 | Improved |
| apply sin() to weights | CNN · SAR interferograms | MSE: 0.0200 -> 0.0120 | Improved |
| use sine instead of cosine basis (ablation) | Sin-CNN · Synthetic attosecond spectrograms | MSE: 0.0090 -> 0.0080 | Improved |
| use sine instead of random Fourier basis (ablation) | Sin-CNN · Synthetic attosecond spectrograms | MSE: 0.0100 -> 0.0080 | Improved |
| make amplitude, frequency, phase learnable (ablation) | Sin-Basis · Synthetic attosecond spectrograms | MSE: 0.0080 -> 0.0060 | Improved |
| apply sin() to weights | CNN · Audio event classification | accuracy: 78.2 -> 84.7 | Improved |
| apply sin() to weights | ViT · Kinetics periodic action | accuracy: 85.3 -> 91.0 | Improved |
| zero-tune Sin-CNN vs fine-tuned CNN | Sin-CNN · Raman transfer | MSE: 0.014 -> 0.011 | Improved |
| apply sin() to weights | CNN · FGSM eps=0.05 | MSE: 0.065 -> 0.039 | Improved |
| apply sin() to weights | CNN · Training time | seconds/epoch: 12.3 -> 13.8 | Hurt |
Closest papers by meaning
- Reconstructing Randomly Masked Spectra Helps DNNs Identify Discriminant Wavenumbers (TPAMI 2023 · similarity 0.86)
- OpticalNet: An Optical Imaging Dataset and Benchmark Beyond the Diffraction Limit (CVPR 2025 · similarity 0.84)
- Unraveling Metameric Dilemma for Spectral Reconstruction: A High-Fidelity Approach via Semi-Supervised Learning (NeurIPS 2025 · similarity 0.83)
- SONIC: Spectral Oriented Neural Invariant Convolutions (ICLR 2026 · similarity 0.83)
- Spectral and Polarization Vision: Spectro-polarimetric Real-world Dataset (CVPR 2024 · similarity 0.82)
- Exploring Invariance in Images through One-way Wave Equations (ICML 2025 · similarity 0.82)
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.
