Xu Cao 曹旭

Hi there! I am a research scientist in the graphics team at CyberAgent AI Lab. Previously, I worked as a Specially Appointed Assistant Professor (特任助教) in Computer Vision Lab. at Osaka University from 2022 to 2023.

I earned my Ph.D. (Apr. 2019 - Mar. 2022) from Osaka University under the supervision of Prof. Yasuyuki Matsushita. Prior to that, I received M.S. degree from Nagoya University in Mar. 2019 and B.S. degree from Nanjing University of Aeronautics and Astronautics (NUAA) in Jun. 2016.

My research interest includes 3D and physics-based vision.

Email  |  Google Scholar  |  Github  |  LinkedIn  |  Twitter

profile photo

"It can scarcely be denied that the supreme goal of all theory is to make the irreducible basic elements as simple and as few as possible without having to surrender the adequate representation of a single datum of experience." (Albert Einstein, 1933)


Highlighted are my favorite works.

SuperNormal: Neural Surface Reconstruction via Multi-View Normal Integration
Xu Cao, Takafumi Taketomi
CVPR 2024
arXiv  |  benchmark results  |  code [SuperNormal]  |  code [Neuralangelo with DFD]

SuperNormal efficiently recovers high-fidelity shapes using multi-view normal maps estimated by photometric stereo. We accelerate neural SDF training by directional finite difference (DFD) and patch-based sampling.

Multi-View Azimuth Stereo via Tangent Space Consistency (MVAS via TSC)
Xu Cao, Hiroaki Santo, Fumio Okura, Yasuyuki Matsushita
CVPR 2023
project page  |  arXiv  |  code

We find a strictly held consistency in multi-view stereo when using azimuth maps as inputs and leverage this consistency for neural SDF optimization.

ECON: Explicit Clothed humans Optimized via Normal integration
Yuliang Xiu, Jinlong Yang, Xu Cao, Dimitrios Tzionas, Michael J. Black
CVPR 2023 (Highlight)
project page  |  arXiv  |  code  |  video

ECON is designed for full-body human digitization from a single-view color image and features robust performance on in-the-wild images with loose clothing or challenging poses.

Bilateral Normal Integration (BiNI)
Xu Cao, Hiroaki Santo, Boxin Shi, Fumio Okura, Yasuyuki Matsushita
ECCV 2022
PDF  |  code

BiNI is a variational approach for recovering the surface from a single-view normal map, which is capable of preserving depth discontinuities.

Shape and Albedo Recovery by Your Phone using Stereoscopic Flash and No-Flash Photography
Xu Cao, Michael Waechter, Boxin Shi, Ye Gao, Bo Zheng, Fumio Okura, Yasuyuki Matsushita
IJCV 2022

Using a stereo camera and a flashlight that are common on modern mobile phones can recover high-fidelity surfaces.

Normal Integration via Inverse Plane Fitting with Minimum Point-to-Plane Distance
Xu Cao, Boxin Shi, Fumio Okura, Yasuyuki Matsushita
CVPR 2021
PDF  |  video  |  code

Plane fitting is useful for normal estimation; plane fitting is now useful for normal integration.

Stereoscopic Flash and No-Flash Photography for Shape and Albedo Recovery
Xu Cao, Michael Waechter, Boxin Shi, Ye Gao, Bo Zheng, Yasuyuki Matsushita
CVPR 2020
PDF  |  video  |  code

Flashed stereo camera, recovered fine shapes.

PSNet: A Style Transfer Network for Point Cloud Stylization on Geometry and Color
Xu Cao, Weimin Wang, Katashi Nagao, Ryosuke Nakamura
WACV 2020
PDF  |  code

Learnt features from PointNet can stylize point clouds.

Website adapted from Jon Barron.