Publication

InstantGeoAvatar: Effective geometry and appearance modeling of animatable avatars from monocular video

Conference Article

Conference

Asian Conference on Computer Vision (ACCV)

Edition

17th

Pages

255-277

Doc link

http://dx.doi.org/10.1007/978-981-96-0960-4_16

File

Download the digital copy of the doc pdf document

Abstract

We present InstantGeoAvatar, a method for efficient and effective learning from monocular video of detailed 3D geometry and appearance of animatable implicit human avatars. Our key observation is that the optimization of a hash grid encoding to represent a signed distance function (SDF) of the human subject is fraught with instabilities and bad local minima. We thus propose a principled geometry-aware SDF regularization scheme that seamlessly fits into the volume rendering pipeline and adds negligible computational overhead. Our regularization scheme significantly outperforms previous approaches for training SDFs on hash grids. We obtain competitive results in geometry reconstruction and novel view synthesis in as little as five minutes of training time, a significant reduction from the several hours required by previous work. InstantGeoAvatar represents a significant leap forward towards achieving interactive reconstruction of virtual avatars.

Categories

computer vision.

Author keywords

3D Computer Vision, Human Avatars, Neural Radiance Fields, Clothed Human Modeling

Scientific reference

A.F. Budria, A. López, O. Lorente and F. Moreno-Noguer. InstantGeoAvatar: Effective geometry and appearance modeling of animatable avatars from monocular video, 17th Asian Conference on Computer Vision, 2024, Hanoi, in Computer Vision – ACCV 2024, Vol 15472 of Lecture Notes in Computer Science, pp. 255-277, 2024.