평범한 필기장

Paper : https://arxiv.org/abs/2303.04761 Video-P2P: Video Editing with Cross-attention ControlThis paper presents Video-P2P, a novel framework for real-world video editing with cross-attention control. While attention control has proven effective for image editing with pre-trained image generation models, there are currently no large-scale video gearxiv.orgGithub : https://github.com/dvlab-resea..

0. AbstractKey Challenge : Naive DDIM inversion process의 각 step에서의 randomness와 inaccuracy에 의해 발생하는 error를 제한하는 것.이는 video editing에서 temporal inconsistency를 야기할 수 있다.1. Introduction 본 논문은 diffusion model을 이용해서 zero-shot video editing method를 만드는 것을 목표로 한다. Inversion process는 temporally cohorent initial latents의 sequence를 제공함으로써 video editing 결과에 도움을 준다. 그러나 아래의 이미지처럼 direct inversion process는 pot..

Paper : https://arxiv.org/abs/2403.12002 DreamMotion: Space-Time Self-Similar Score Distillation for Zero-Shot Video EditingText-driven diffusion-based video editing presents a unique challenge not encountered in image editing literature: establishing real-world motion. Unlike existing video editing approaches, here we focus on score distillation sampling to circumvent the stanarxiv.orgProject P..

Paper : https://arxiv.org/abs/2311.17338 MagDiff: Multi-Alignment Diffusion for High-Fidelity Video Generation and EditingThe diffusion model is widely leveraged for either video generation or video editing. As each field has its task-specific problems, it is difficult to merely develop a single diffusion for completing both tasks simultaneously. Video diffusion sorely relyinarxiv.org1. Introduc..

Paper : https://arxiv.org/abs/2403.14617v3 Videoshop: Localized Semantic Video Editing with Noise-Extrapolated Diffusion InversionWe introduce Videoshop, a training-free video editing algorithm for localized semantic edits. Videoshop allows users to use any editing software, including Photoshop and generative inpainting, to modify the first frame; it automatically propagates those charxiv.org0. ..