IMPUS: Image Morphing with Perceptually-Uniform Sampling Using Diffusion Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yang, Zhaoyuan, Yu, Zhengyang, Xu, Zhiwei, Singh, Jaskirat, Zhang, Jing, Campbell, Dylan, Tu, Peter, Hartley, Richard
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916160702775296
author Yang, Zhaoyuan
Yu, Zhengyang
Xu, Zhiwei
Singh, Jaskirat
Zhang, Jing
Campbell, Dylan
Tu, Peter
Hartley, Richard
author_facet Yang, Zhaoyuan
Yu, Zhengyang
Xu, Zhiwei
Singh, Jaskirat
Zhang, Jing
Campbell, Dylan
Tu, Peter
Hartley, Richard
contents We present a diffusion-based image morphing approach with perceptually-uniform sampling (IMPUS) that produces smooth, direct and realistic interpolations given an image pair. The embeddings of two images may lie on distinct conditioned distributions of a latent diffusion model, especially when they have significant semantic difference. To bridge this gap, we interpolate in the locally linear and continuous text embedding space and Gaussian latent space. We first optimize the endpoint text embeddings and then map the images to the latent space using a probability flow ODE. Unlike existing work that takes an indirect morphing path, we show that the model adaptation yields a direct path and suppresses ghosting artifacts in the interpolated images. To achieve this, we propose a heuristic bottleneck constraint based on a novel relative perceptual path diversity score that automatically controls the bottleneck size and balances the diversity along the path with its directness. We also propose a perceptually-uniform sampling technique that enables visually smooth changes between the interpolated images. Extensive experiments validate that our IMPUS can achieve smooth, direct, and realistic image morphing and is adaptable to several other generative tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2311_06792
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle IMPUS: Image Morphing with Perceptually-Uniform Sampling Using Diffusion Models
Yang, Zhaoyuan
Yu, Zhengyang
Xu, Zhiwei
Singh, Jaskirat
Zhang, Jing
Campbell, Dylan
Tu, Peter
Hartley, Richard
Computer Vision and Pattern Recognition
We present a diffusion-based image morphing approach with perceptually-uniform sampling (IMPUS) that produces smooth, direct and realistic interpolations given an image pair. The embeddings of two images may lie on distinct conditioned distributions of a latent diffusion model, especially when they have significant semantic difference. To bridge this gap, we interpolate in the locally linear and continuous text embedding space and Gaussian latent space. We first optimize the endpoint text embeddings and then map the images to the latent space using a probability flow ODE. Unlike existing work that takes an indirect morphing path, we show that the model adaptation yields a direct path and suppresses ghosting artifacts in the interpolated images. To achieve this, we propose a heuristic bottleneck constraint based on a novel relative perceptual path diversity score that automatically controls the bottleneck size and balances the diversity along the path with its directness. We also propose a perceptually-uniform sampling technique that enables visually smooth changes between the interpolated images. Extensive experiments validate that our IMPUS can achieve smooth, direct, and realistic image morphing and is adaptable to several other generative tasks.
title IMPUS: Image Morphing with Perceptually-Uniform Sampling Using Diffusion Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2311.06792