# Learning to Generate Conditional Tri-plane for 3D-aware Expression Controllable Portrait Animation

Source: [Learning to Generate Conditional Tri-plane for 3D-aware Expression Controllable Portrait Animation](<https://www.sota2.com/research/paper/68fa743f1e209131cea607fd>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-06T05:00:05.533Z

Authors: Taekyung Ki, Dongchan Min, Gyeongsu Chae

Published: 2024-03-31T10:13:55.000Z

## Abstract

In this paper, we present Export3D, a one-shot 3D-aware portrait animation method that is able to control the facial expression and camera view of a given portrait image. To achieve this, we introduce a tri-plane generator with an effective expression conditioning method, which directly generates a tri-plane of 3D prior by transferring the expression parameter of 3DMM into the source image. The tri-plane is then decoded into the image of different view through a differentiable volume rendering. Existing portrait animation methods heavily rely on image warping to transfer the expression in the motion space, challenging on disentanglement of appearance and expression. In contrast, we propose a contrastive pre-training framework for appearance-free expression parameter, eliminating undesirable appearance swap when transferring a cross-identity expression. Extensive experiments show that our pre-training framework can learn the appearance-free expression representation hidden in 3DMM, and our model can generate 3D-aware expression controllable portrait images without appearance swap in the cross-identity manner.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [3D Portrait Animation (Cross Reenactment) on VFHQ 1.0 (test)](<https://www.sota2.com/research/sota/3d-portrait-animation-cross-reenactment-on-vfhq-1-0-test.md>): 3D Portrait Animation (Cross Reenactment); VFHQ 1.0 (test); CSIM: 69.4.

- [Portrait Animation on VFHQ Same-identity](<https://www.sota2.com/research/sota/portrait-animation-on-vfhq-same-identity.md>): Portrait Animation; VFHQ Same-identity; PSNR: 23.555.

- [Portrait Animation on TalkingHead-1KH Same-identity](<https://www.sota2.com/research/sota/portrait-animation-on-talkinghead-1kh-same-identity.md>): Portrait Animation; TalkingHead-1KH Same-identity; PSNR: 23.239.

- [Portrait Animation on TalkingHead-1KH Cross-identity](<https://www.sota2.com/research/sota/portrait-animation-on-talkinghead-1kh-cross-identity.md>): Portrait Animation; TalkingHead-1KH Cross-identity; CSIM: 0.772.

- [Portrait Animation on VFHQ Same-identity with background (test)](<https://www.sota2.com/research/sota/portrait-animation-on-vfhq-same-identity-with-background-test.md>): Portrait Animation; VFHQ Same-identity with background (test); PSNR: 23.555.

- [Portrait Animation on VFHQ (Cross-identity)](<https://www.sota2.com/research/sota/portrait-animation-on-vfhq-cross-identity.md>): Portrait Animation; VFHQ (Cross-identity); CSIM: 69.4.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2404.00636>)

[PDF](<https://arxiv.org/pdf/2404.00636.pdf>)
