Research map / Image generation / Fast diffusion sampling
Image Generation using Continuous Filter Atoms
Ze Wang, Seunghyun Hwang, Zichen Miao, Qiang Qiu
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 |
|---|---|---|---|
| plug continuous filter atoms into Pix2Pix | Pix2Pix · RC49 | FID: 251.49 -> 195.01 | Improved |
| plug continuous filter atoms into Pix2Pix | Pix2Pix · RCA20 | FID: 218.88 -> 85.06 | Improved |
| plug continuous filter atoms into Pix2Pix | Pix2Pix · RT20 | FID: 206.49 -> 70.92 | Improved |
| plug continuous filter atoms into Pix2Pix | Pix2Pix · RA20 | LPIPS diversity: - -> 0.068 | Improved |
| plug continuous filter atoms into CycleGAN | CycleGAN · Cell200 | FID: 30.44 -> 22.22 | Improved |
| plug continuous filter atoms into CycleGAN | CycleGAN · SteeringAngle | FID: 73.03 -> 52.39 | Improved |
| plug continuous filter atoms into CycleGAN | CycleGAN · CACD2000 | FID: 58.43 -> 57.15 | No clear effect |
| plug continuous filter atoms into CycleGAN | CycleGAN · CACD2000 | LPIPS diversity: - -> 0.0018 | No clear effect |
| replace BasisGAN atom generator with neural ODE atoms | BasisGAN · Edge->Handbag | FID: 88.76 -> 38.07 | Improved |
| replace BasisGAN atom generator with neural ODE atoms | BasisGAN · Map->Satellite | FID: 35.54 -> 98.72 | Hurt |
| replace BasisGAN atom generator with neural ODE atoms | BasisGAN · Night->Day | LPIPS diversity: 0.184 -> 0.264 | Improved |
| replace BasisGAN atom generator with neural ODE atoms | BasisGAN · Edge->Handbag | LPIPS diversity: 0.350 -> 0.276 | Hurt |
Closest papers by meaning
- A Simple Approach to Unifying Diffusion-based Conditional Generation (ICLR 2025 · similarity 0.84)
- Large Convolutional Model Tuning via Filter Subspace (ICLR 2025 · similarity 0.84)
- Continuity-Preserving Convolutional Autoencoders for Learning Continuous Latent Dynamical Models from Images (ICLR 2025 · similarity 0.84)
- Rethinking Peculiar Images by Diffusion Models: Revealing Local Minima’s Role (AAAI 2024 · similarity 0.83)
- HART: Efficient Visual Generation with Hybrid Autoregressive Transformer (ICLR 2025 · similarity 0.83)
- NoiseCtrl: A Sampling-Algorithm-Agnostic Conditional Generation Method for Diffusion Models (CVPR 2025 · similarity 0.83)
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.
