본문 바로가기

전체 글92

One-Minute Video Generation with Test-Time Training 깃허브 링크 : https://github.com/time-to-move/TTM GitHub - time-to-move/TTM: Official Pytorch Implementation for "Time-to-Move: Training-Free Motion Controlled Video Generation vOfficial Pytorch Implementation for "Time-to-Move: Training-Free Motion Controlled Video Generation via Dual-Clock Denoising" - time-to-move/TTMgithub.com 논문 링크 : https://arxiv.org/abs/2511.08633 Time-to-Move: Training-Free M.. 2026. 7. 24.
VACE : All-in-One Video Creation and Editing 깃허브 링크 : https://github.com/ali-vilab/VACE GitHub - ali-vilab/VACE: [ICCV 2025] Official implementations for paper: VACE: All-in-One Video Creation and Editing[ICCV 2025] Official implementations for paper: VACE: All-in-One Video Creation and Editing - ali-vilab/VACEgithub.com 논문 링크 : https://arxiv.org/abs/2503.07598 VACE: All-in-One Video Creation and EditingDiffusion Transformer has demonstrat.. 2026. 7. 24.
[3D] 3차원 좌표계 변환, World 좌표 → Camera 좌표 12.1 3차원 기하와 캘리브레이션Calibrataion : 어떤 좌표계 상의 점 → 다른 좌표계의 점 1) 세계 좌표계와 카메라 좌표계카메라는 세계 좌표계를 기준으로 이동 $\textbf{t}=\left(t_x,t_y,t_z\right)$ 과 회전 $ r_1, r_2, \ldots, r_9 $ 진행 세계 좌표계를 카메라 좌표계로 변환 :$ (x_w,y_w,z_w)\to(x_c,y_c,z_c) $ 회전과 이동이 작용하는 것을 나타내기 위해 보통 3 x 4 행렬 $(\textbf{R}|\textbf{t})$ 사용 (외부 행렬 extrinsic matrix) $\begin{pmatrix}x_c\\y_c\\z_c\end{pmatrix}=\begin{pmatrix}r_1&r_2&r_3&t_x\\r_4&r_.. 2025. 12. 5.
[논문리뷰] Video Depth Anything : Consistent DepthEestimation for Super-Long Videos 깃허브 : https://github.com/DepthAnything/Video-Depth-Anything GitHub - DepthAnything/Video-Depth-Anything: [CVPR 2025 Highlight] Video Depth Anything: Consistent Depth Estimation for Super-L[CVPR 2025 Highlight] Video Depth Anything: Consistent Depth Estimation for Super-Long Videos - DepthAnything/Video-Depth-Anythinggithub.com논문 링크 : https://arxiv.org/pdf/2501.12375 [발표 PPT] 2025. 10. 20.
[논문 리뷰] GPEN : GAN Prior Embedded Network for Blind Face Restoration in the Wild 깃허브 링크 : https://github.com/yangxy/GPEN GitHub - yangxy/GPENContribute to yangxy/GPEN development by creating an account on GitHub.github.com 논문 링크 : https://arxiv.org/pdf/2105.06070 1. IntroductionBlind Face Restoration (BFR) Task를 다룸 실제 환경에선 열화가 다양하기 때문에 BFR Task는 매우 어려운 문제임GAN의 발전 이후 극도로 낮은 해상도의 입력을 재구성하려는 시도가 생김- 사전 학습된 GAN을 변경하지 않아 복원 품질이 불안정한 문제 발생 본 논문은 GAN과 DNN의 장점을 통합함GAN을 사전 학습한 뒤, 얼굴.. 2025. 9. 28.
[논문리뷰] CodeFormer : Towards Robust Blind Face Restoration with Codebook Lookup Transformer 깃허브 : https://github.com/sczhou/CodeFormer GitHub - sczhou/CodeFormer: [NeurIPS 2022] Towards Robust Blind Face Restoration with Codebook Lookup Transformer[NeurIPS 2022] Towards Robust Blind Face Restoration with Codebook Lookup Transformer - sczhou/CodeFormergithub.com 논문 링크 : https://proceedings.neurips.cc/paper_files/paper/2022/file/c573258c38d0a3919d8c1364053c45df-Paper-Conference.pdf [발표 P.. 2025. 1. 20.