🔥 Core
- Denoising Diffusion Probabilistic Models(DDPM) (paper), (mycode)
- Denoising Diffusion Implicit Models(DDIM) (paper)
- Improved Denoising Diffusion Probabilistic Models(Improved DDPM) (paper)
- Score-Based Generative Modeling through Stochastic Differential Equations (paper)
- Cold Diffusion: Inverting Arbitrary Image Transforms Without Noise (paper)
- Classifier-free Diffusion Guidance (paper)
- Diffusion Models Beat GANs on Image Synthesis (paper)
- Scalable Diffusion Models with Transformers (paper)
- High-Resolution Image Synthesis with Latent Diffusion Models(Stable Diffusion) (paper)
🔥 Segmentation
- SegDiff (paper)
- MedSegDiff (paper)
- SegDiff (paper)
- Diffusion Models for Implicit Image Segmentation Ensembles (paper)
- DiffusionDet (paper)
- DiffusionInst (paper)
- Label-Efficient semantic segmentation with Diffusion Models (paper)
- Accelerating Diffusion Models via Pre-segmentation Diffusion Sampling for Medical Image Segmentation (paper)
🔥 Need to Read
- Large Scale GAN Training for High Fidelity Natural Image Synthesis(BigGAN) (paper)
- A Style-Based Generator Architecture for Generative Adversarial Networks(StyleGAN) (paper)
- Analyzing and Improving the Image Quality of StyleGAN(StyleGAN2) (paper)
- CLIP : Learning Transferable Visual Models From Natural Language Supervision (paper)
- Make Custom Dataset Class
- Try My custom U-Net with concat Image(Diff input :
mask
| Image :guide
)|fail|
- Adjust MedSegdiff or Ensemble Paper's method to concat Image
- Mix transformer
- Do various experiments
- Make recognize instance model