paper-with-me

홈 › Papers

Improved Video VAE for Latent Video Diffusion Model

2024-11-10 · CVPR 2025 1 · Pingyu Wu, Kai Zhu, Yu Liu, Liming Zhao, Wei Zhai, Yang Cao, Zheng-Jun Zha

Variational Autoencoder (VAE) aims to compress pixel data into low-dimensional latent space, playing an important role in OpenAI's Sora and other latent video diffusion generation models. While most of existing video VAEs inflate a pretrained image VAE into the 3D causal structure for temporal-spatial compression, this paper presents two astonishing findings: (1) The initialization from a well-trained image VAE with the same latent dimensions suppresses the improvement of subsequent temporal compression capabilities. (2) The adoption of causal reasoning leads to unequal information interactions and unbalanced performance between frames. To alleviate these problems, we propose a keyframe-based temporal compression (KTC) architecture and a group causal convolution (GCConv) module to further improve video VAE (IV-VAE). Specifically, the KTC architecture divides the latent space into two branches, in which one half completely inherits the compression prior of keyframes from a lower-dimension image VAE while the other half involves temporal compression through 3D group causal convolution, reducing temporal-spatial conflicts and accelerating the convergence speed of video VAE. The GCConv in above 3D half uses standard convolution within each frame group to ensure inter-frame equivalence, and employs causal logical padding between groups to maintain flexibility in processing variable frame video. Extensive experiments on five benchmarks demonstrate the SOTA video reconstruction and generation capabilities of the proposed IV-VAE (https://wpy1999.github.io/IV-VAE/).

📄 PDF Abstract BibTeX arXiv:2411.06449

Code (0)

등록된 구현이 없습니다.

Tasks

modelVideo Reconstruction

Methods 이 논문이 사용한 방법론

Diffusion Diffusion models generate samples by gradually removing noise from a signal, and their training objective can be expressed as a reweighted variational lower-bound…
Causal Convolution Causal convolutions are a type of convolution used for temporal data which ensures the model cannot violate the ordering in…
Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…
SPEED The monocular depth estimation (MDE) is the task of estimating depth from a single frame. This information is an essential knowledge in many computer vision tasks such as scene…

Similar Papers 제목 키워드 기반

Noise Crystallization and Liquid Noise: Zero-shot Video Generation using Image Diffusion Models

2024-10-05 · Muhammad Haaris Khan, Hadrien Reynaud, Bernhard Kainz

Although powerful for image generation, consistent and controllable video is a longstanding problem for diffusion models. Video models require extensive training and computational resources, leading to high costs and lar…

Image GenerationStyle TransferVideo GenerationVideo Style Transfer

Brick-Diffusion: Generating Long Videos with Brick-to-Wall Denoising

2025-01-06 · Yunlong Yuan, Yuanfan Guo, Chunwei Wang, Hang Xu 외

Recent advances in diffusion models have greatly improved text-driven video generation. However, training models for long video generation demands significant computational power and extensive data, leading most video di…

DenoisingVideo Generation

Denoising Reuse: Exploiting Inter-frame Motion Consistency for Efficient Video Latent Generation

2024-09-19 · Chenyu Wang, Shuo Yan, Yixuan Chen, Yujiang Wang 외

Video generation using diffusion-based models is constrained by high computational costs due to the frame-wise iterative diffusion process. This work presents a Diffusion Reuse MOtion (Dr. Mo) network to accelerate laten…

DenoisingVideo Generation

SV3D: Novel Multi-view Synthesis and 3D Generation from a Single Image using Latent Video Diffusion

2024-03-18 · Vikram Voleti, Chun-Han Yao, Mark Boss, Adam Letts 외

We present Stable Video 3D (SV3D) -- a latent video diffusion model for high-resolution, image-to-multi-view generation of orbital videos around a 3D object. Recent work on 3D generation propose techniques to adapt 2D ge…

3D Generation3D ReconstructionImage to 3DNovel View Synthesis

HorizonRelight: Relighting Long-horizon Videos Consistently via Diffusion Transformers

2026-06-27 · Jing Yang, Mayoore Jaiswal, Zian Wang, Steven Zeng 외 arxiv

Diffusion-based video relighting enables controllable relighting from a single input video, but modern video diffusion backbones are trained on short clips and applied to long-horizon videos through chunked sliding-windo…