Notice
Recent Posts
Recent Comments
Link
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 | 31 |
Tags
- VQ-diffusion
- 베이지안 정리
- VQ-VAE
- learning to prompt for continual learning
- mmcv
- DualPrompt
- Face Pose Estimation
- Face Alignment
- Continual Learning
- state_dict()
- Class Incremental
- Class Incremental Learning
- Markov transition matrix
- Img2pose
- CIL
- Facial Landmark Localization
- timm
- Vector Quantized Diffusion Model for Text-to-Image Synthesis
- CVPR2022
- ENERGY-BASED MODELS FOR CONTINUAL LEARNING
- requires_grad
- img2pose: Face Alignment and Detection via 6DoF
- Mask diffusion
- Mask-and-replace diffusion strategy
- L2P
- prompt learning
- Energy-based model
- learning to prompt
- Discrete diffusion
- PnP algorithm
Archives
- Today
- Total
Computer Vision , AI
[One-page summary] ObjectStitch: Generative Object Compositing (CVPR 2023) by Song et al. 본문
Paper_review[short]
[One-page summary] ObjectStitch: Generative Object Compositing (CVPR 2023) by Song et al.
Elune001 2024. 1. 16. 01:31● Summary: a single framework for Image composition (color harmonization, geometric correction, shadow generation) with no label
● Approach highlight
- Self-supervised learning: segment object from the original image and mask that portion
- Content adaptor for object identity: image-to-text embedding using CLIP embedding (to use a diffusion model designed for text embedding)

- Diffusion with the masked image

● Main Results



● Discussion
- Wouldn't artifacts appear at the boundary of the Mask area?
- How would you generate shadows outside the Mask area?