Research map / Learning theory and algorithms / Combinatorial and search optimization
A Flexible Convolutional Solver for Fast Style Transfers
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 |
|---|---|---|---|
| Increase the number of learned updates from N:1 to N:4 (ablation) | Flexible convolutional solver · MS-COCO validation content-style pairs | LTV: 0.33 ± 0.07 -> 0.24 ± 0.05 | Helped |
| Increase channel count from L:16 to L:64 at N:3 (ablation) | Flexible convolutional solver · MS-COCO validation content-style pairs | LTV: 0.29 ± 0.07 -> 0.26 ± 0.06 | Helped |
| Use backward-map subgradient h instead of ReLU (ablation) | Flexible convolutional solver · MS-COCO validation content-style pairs | LTV: 0.27 ± 0.06 -> 0.26 ± 0.06 | Helped |
| Add locally affine matting-Laplacian regularization at runtime | Flexible convolutional solver · Photo style transfer | computation time: 0.79 -> 0.44; 3.12 -> 1.86; 15.23 -> 8.11 | Helped |
| Add short-term temporal consistency regularization at runtime | Flexible convolutional solver · Video style transfer | temporal consistency | Helped |
| Add a user-provided reference texture to guide super-resolution (ablation) | Flexible convolutional solver · Texture super-resolution, example 1 | SNR Gram: 6.26 dB -> 16.5 dB | Helped |
| Add a user-provided reference texture to guide super-resolution (ablation) | Flexible convolutional solver · Texture super-resolution, example 2 | SNR Gram: 8.67 dB -> 10.9 dB | Helped |
| Add a user-provided reference texture to guide super-resolution (ablation) | Flexible convolutional solver · Texture super-resolution, example 1 | PSNR: 28.7 dB -> 26.5 dB | Hurt |
| Add a user-provided reference texture to guide super-resolution (ablation) | Flexible convolutional solver · Texture super-resolution, example 2 | PSNR: 32.6 dB -> 32.1 dB | Hurt |
| Use graph filtering to compute the matting-Laplacian proximal operator | Flexible convolutional solver · Photo style transfer 1024 × 512 without Laplacian | computation time: 0.34 -> 0.37 | Hurt |
Closest papers by meaning
- Multimodal Transfer: A Hierarchical Deep Convolutional Neural Network for Fast Artistic Style Transfer (CVPR 2017 · similarity 0.93)
- Real-Time Neural Style Transfer for Videos (CVPR 2017 · similarity 0.91)
- Image Style Transfer Using Convolutional Neural Networks (CVPR 2016 · similarity 0.91)
- Adaptive Convolutions for Structure-Aware Style Transfer (CVPR 2021 · similarity 0.90)
- Puff-Net: Efficient Style Transfer with Pure Content and Style Feature Fusion Network (CVPR 2024 · similarity 0.88)
- ArtBank: Artistic Style Transfer with Pre-trained Diffusion Model and Implicit Style Prompt Bank (AAAI 2024 · 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.