| 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
- diffpir
- ilvr
- focal detection network
- f2dnet
- neighboring pixel relationships
- Ai
- linear inverse problem
- ddrm
- object detection
- face forgery detection
- facenet
- unconditional generative models
- stochastic contraction theory
- IR
- daod
- fourmer
- ddim
- unconditional ddpm
- pedestrian detection
- ccdf
- ddpm
- deepfake detection
- ML
- hqs-algorithm
- semi-supervied learning
- image restoration
- ssda-yolo
- conditional diffusion
- diffusion
- Triplet
Archives
- Today
- Total
목록semi-supervied learning (1)
Stand on the shoulders of giants
[paper review] SSDA-YOLO: Semi-supervised domain adaptive YOLO for cross-domain object detection
Introduction해당 논문에서는 one-stage detector인 YOLOv5와 domain adaptiation을 결합해 cross-domain detection 성능을 개선할 수 있는 semi-supervised domain adaptive YOLO를 제안했다.ContributionsDAOD 문제를 해결할 수 있는 novel한 SSDA-YOLO를 제안함Penalty function인 distillation loss, consistency loss를 새롭게 제안해 성능을 개선시킴.저명한domain transfer 실험들에서 성능 개선Outdated detector인 Faster R-CNN과 비교했을때 발전된 Detector가 DAOD 필드에 필요함을 보임.Related Works1) Object..
Paper reviews
2024. 7. 18. 22:50