Research map / Image generation / Editing and style transfer
Neural Stereoscopic Image Style Transfer
Xinyu Gong, Haozhi Huang, Lin Ma, Fumin Shen, Wei Liu, Tong Zhang
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 |
|---|---|---|---|
| train with image-level view loss (ablation) | SingleImage · Driving, 4 styles | MVL: 2033 -> 1121 | Improved |
| add feature aggregation block (warp, gate, concat) (ablation) | SingleImage-IV · Driving, 4 styles | MVL: 1121 -> 1028 | Improved |
| add feature-level view loss to image-level view loss (ablation) | Stereo-FA-IV · Driving, 4 styles | MVL: 1028 -> 1014 | Improved |
| add feature-level view loss to image-level view loss (ablation) | Stereo-FA-IV · Driving, 4 styles | MCL: 481056 -> 445336 | Improved |
| add feature-level view loss to image-level view loss (ablation) | Stereo-FA-IV · Driving, 4 styles | MSL: 410 -> 417 | Hurt |
| fine-tune whole network instead of freezing disparity sub-network (ablation) | Stereo-FA-IV · Driving, 4 styles | MVL: 1028 -> 1022 | No clear effect |
| concatenate right-view features with left (ablation) | SingleImage-IV · Feature aggregation ablation | MVL: 1121 -> 1068 | Improved |
| add warping and gating before concatenation (ablation) | CON-IV · Feature aggregation ablation | MVL: 1068 -> 1028 | Improved |
| add warping and gating before concatenation (ablation) | CON-IV · Feature aggregation ablation | MCL: 489555 -> 481056 | Improved |
| add warping and gating before concatenation (ablation) | CON-IV · Feature aggregation ablation | MSL: 328 -> 410 | Hurt |
| use stereo dual-path method instead of single image method | Johnson et al. · Waterloo-IVC user study | votes: 143 vs 29 (Candy) | Improved |
| use stereo dual-path method instead of single image method | Johnson et al. · Waterloo-IVC user study (Fish) | votes: 166 vs 14 | Improved |
Closest papers by meaning
- Stereoscopic Neural Style Transfer (CVPR 2018 · similarity 0.94)
- Joint Bilateral Learning for Real-time Universal Photorealistic Style Transfer (ECCV 2020 · similarity 0.88)
- Real-Time Neural Style Transfer for Videos (CVPR 2017 · similarity 0.88)
- MetaStyle: Three-Way Trade-off among Speed, Flexibility, and Quality in Neural Style Transfer (AAAI 2019 · similarity 0.88)
- A Geometrical Perspective on Image Style Transfer with Adversarial Learning (TPAMI 2020 · similarity 0.87)
- Artistic Style Transfer for Videos and Spherical Images (IJCV 2018 · 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.
