Research map / Image generation / Editing and style transfer
Neural Collage Transfer: Artistic Reconstruction via Material Manipulation
Ganghun Lee, Minji Kim, Yunsu Lee, Minsu Lee, Byoung-Tak 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 |
|---|---|---|---|
| use model-based SAC instead of model-free SAC (ablation) | collage agent · Scene target, DTD materials, 64x64 | MSE: 0.042 -> 0.031 | Improved |
| use SAC instead of DDPG in model-based agent (ablation) | collage agent · Scene target, DTD materials, 64x64 | MSE: 0.039 -> 0.031 | Improved |
| use SAC instead of DDPG in model-free agent (ablation) | collage agent · Scene target, DTD materials, 64x64 | MSE: 0.305 -> 0.042 | Improved |
| use a learned WGAN-GP discriminator reward instead of MSE reward (ablation) | collage agent · Scene target, DTD materials, 64x64 | MSE: 0.122 -> 0.031 | Improved |
| use WGAN-GP instead of SNGAN discriminator for reward (ablation) | collage agent · Scene target, DTD materials, 64x64 | MSE: 0.040 -> 0.031 | Improved |
| use WGAN-GP instead of GNGAN discriminator for reward (ablation) | collage agent · Scene target, DTD materials, 64x64 | MSE: 0.084 -> 0.031 | Improved |
| increase the multi-scale collage scale from 32 to 128 (ablation) | multi-scale collage · CLIP score content | CLIP score content: 0.280 -> 0.225 | Hurt |
| increase the multi-scale collage scale from 32 to 128 (ablation) | multi-scale collage · CLIP score collage | CLIP vote collage: 0.100 -> 1.000 | Improved |
| increase the number of scraps from 10 to 30 (ablation) | collage agent · Scene target, DTD materials | MSE | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- StyTR-2 — StyTr2: Image Style Transfer With Transformers (CVPR 2022) · better on 2 of 2 shared results
- AdaAttn — AdaAttN: Revisit Attention Mechanism in Arbitrary Neural Style Transfer (ICCV 2021) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- SoftCollage: A Differentiable Probabilistic Tree Generator for Image Collage (CVPR 2022 · similarity 0.87)
- Paint Transformer: Feed Forward Neural Painting With Stroke Prediction (ICCV 2021 · similarity 0.85)
- SCORE: Semantic Collage by Optimizing Rendered Elements (AAAI 2026 · similarity 0.85)
- MatFuse: Controllable Material Generation with Diffusion Models (CVPR 2024 · similarity 0.85)
- ArtBank: Artistic Style Transfer with Pre-trained Diffusion Model and Implicit Style Prompt Bank (AAAI 2024 · similarity 0.85)
- PrefPaint: Aligning Image Inpainting Diffusion Model with Human Preference (NeurIPS 2024 · similarity 0.84)
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.
