| CARVIEW |
Select Language
HTTP/2 301
server: GitHub.com
content-type: text/html
location: https://hyunsoocha.github.io/perse/
access-control-allow-origin: *
expires: Sun, 28 Dec 2025 09:39:55 GMT
cache-control: max-age=600
x-proxy-cache: MISS
x-github-request-id: 4440:3ABDEF:766AB1:84D1D4:6950F892
accept-ranges: bytes
age: 0
date: Sun, 28 Dec 2025 09:29:55 GMT
via: 1.1 varnish
x-served-by: cache-bom-vanm7210090-BOM
x-cache: MISS
x-cache-hits: 0
x-timer: S1766914195.059743,VS0,VE212
vary: Accept-Encoding
x-fastly-request-id: 4c016f81490b2c41fbd33ea2f6680fcdc5163019
content-length: 162
HTTP/2 200
server: GitHub.com
content-type: text/html; charset=utf-8
last-modified: Mon, 08 Dec 2025 04:58:44 GMT
access-control-allow-origin: *
etag: W/"69365b04-7a6e"
expires: Sun, 28 Dec 2025 09:39:55 GMT
cache-control: max-age=600
content-encoding: gzip
x-proxy-cache: MISS
x-github-request-id: 3AE5:292AC1:7723EF:858A00:6950F893
accept-ranges: bytes
age: 0
date: Sun, 28 Dec 2025 09:29:55 GMT
via: 1.1 varnish
x-served-by: cache-bom-vanm7210090-BOM
x-cache: MISS
x-cache-hits: 0
x-timer: S1766914195.285187,VS0,VE214
vary: Accept-Encoding
x-fastly-request-id: 319d0562ef92c8e72a209af0b39087b971c1de6e
content-length: 5239
PERSE: Personalized 3D Generative Avatars from A Single Portrait
PERSE consists of two main components and an application.
Beard A
Beard B
Hair A
Hair B
Hat A
Hat B
Eyebrows A
Eyebrows B
Mouth A
Mouth B
In-The-Wild Image
Latent Space Sample Image
PERSE: Personalized 3D Generative Avatars from A Single Portrait
Seoul National University
CVPR 2025
CVPR 2025
TL;DR: Given a reference portrait image input, PERSE generates an animatable 3D personalized avatar with disentangled and editable control over various facial attributes.
Overview
PERSE consists of two main components and an application.
First, we generate a 2D monocular synthetic dataset from a single
image using our video diffusion model named as
portrait-CHAMP. The synthetic dataset maintains the same
identity as the input image but is edited with different attributes,
forming a 2D monocular video with the same head pose and facial
expression.
Next, we train 3DGS avatar using the 2D synthetic dataset. The 3DGS
avatar is an avatar model conditioned on a latent space, allowing
disentangled control over attributes.
Finally, PERSE avatar model enables attribute transfer from
in-the-wild 2D images, making interpolation possible between the
pretrained latent space and the in-the-wild attributes.
Results
Synthetic Dataset
We generate synthetic dataset consisting of almost a thousand attribute-edited videos using our methods for attribute-edited portrait image generation and animated portrait video generation by portrait-CHAMP
Unseen Pose Rendering
We show rendering results of PERSE avatar model in unseen pose rendering.
Interpolation Between Two Latent Codes
Beard
Beard A
Loading...
Beard B
Hair
Hair A
Loading...
Hair B
Hat
Hat A
Loading...
Hat B
Eyebrows
Eyebrows A
Loading...
Eyebrows B
Mouth
Mouth A
Loading...
Mouth B
We show that PERSE avatar model's interpolated latent codes generate realistic, high-quality avatars with smoothly interpolated attributes.
Facial Attribute Transfer from In-The-Wild Image
We present the results of transferring facial attributes from in-the-wild images. As shown here, the transferred attributes are well rigged to the avatar and are capable of rendering novel poses.
Transferred Rendering in Unseen Pose
We present the results of transferring facial attributes from in-the-wild images.
Interpolation Between Pretrained Latent Code and Novel Latent Code
In-The-Wild Image
Loading...
Latent Space Sample Image
We demonstrate that interpolation between the transferred attribute also produces natural, high-quality avatars.
BibTeX
@InProceedings{Cha_2025_CVPR,
author = {Cha, Hyunsoo and Lee, Inhee and Joo, Hanbyul},
title = {PERSE: Personalized 3D Generative Avatars from A Single Portrait},
booktitle = {Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR)},
month = {June},
year = {2025},
pages = {15953-15962}
}