LatentMan: Generating Consistent Animated Characters using Image Diffusion Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Eldesokey, Abdelrahman, Wonka, Peter
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929368735940608
author Eldesokey, Abdelrahman
Wonka, Peter
author_facet Eldesokey, Abdelrahman
Wonka, Peter
contents We propose a zero-shot approach for generating consistent videos of animated characters based on Text-to-Image (T2I) diffusion models. Existing Text-to-Video (T2V) methods are expensive to train and require large-scale video datasets to produce diverse characters and motions. At the same time, their zero-shot alternatives fail to produce temporally consistent videos with continuous motion. We strive to bridge this gap, and we introduce LatentMan, which leverages existing text-based motion diffusion models to generate diverse continuous motions to guide the T2I model. To boost the temporal consistency, we introduce the Spatial Latent Alignment module that exploits cross-frame dense correspondences that we compute to align the latents of the video frames. Furthermore, we propose Pixel-Wise Guidance to steer the diffusion process in a direction that minimizes visual discrepancies between frames. Our proposed approach outperforms existing zero-shot T2V approaches in generating videos of animated characters in terms of pixel-wise consistency and user preference. Project page https://abdo-eldesokey.github.io/latentman/.
format Preprint
id arxiv_https___arxiv_org_abs_2312_07133
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle LatentMan: Generating Consistent Animated Characters using Image Diffusion Models
Eldesokey, Abdelrahman
Wonka, Peter
Computer Vision and Pattern Recognition
Machine Learning
We propose a zero-shot approach for generating consistent videos of animated characters based on Text-to-Image (T2I) diffusion models. Existing Text-to-Video (T2V) methods are expensive to train and require large-scale video datasets to produce diverse characters and motions. At the same time, their zero-shot alternatives fail to produce temporally consistent videos with continuous motion. We strive to bridge this gap, and we introduce LatentMan, which leverages existing text-based motion diffusion models to generate diverse continuous motions to guide the T2I model. To boost the temporal consistency, we introduce the Spatial Latent Alignment module that exploits cross-frame dense correspondences that we compute to align the latents of the video frames. Furthermore, we propose Pixel-Wise Guidance to steer the diffusion process in a direction that minimizes visual discrepancies between frames. Our proposed approach outperforms existing zero-shot T2V approaches in generating videos of animated characters in terms of pixel-wise consistency and user preference. Project page https://abdo-eldesokey.github.io/latentman/.
title LatentMan: Generating Consistent Animated Characters using Image Diffusion Models
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2312.07133