1 of 294

Interactive Perception & Graphics for

A Universally Accessible Metaverse

Ruofei Du

Senior Research Scientist / Manager

Google AR

www.ruofeidu.com

Twitter: DuRuofei@

me@duruofei.com

2 of 294

Self Intro

www.duruofei.com

3 of 294

Self Intro

Ruofei Du (杜若飞)

4 of 294

Self Intro

Ruofei Du (杜若飞)

5 of 294

Self Intro

Ruofei Du (杜若飞)

Human-Computer Interaction

Geollery CHI '19, Web3D '19, VR '19

Social Street View Web3D '16 Best Paper Award

VideoFields

Web3D '16

SketchyScene

TOG (SIGGRAPH Asia) '19, ECCV '18

Montage4D

I3D '18

JCGT '19

DepthLab UIST '20

13K Installs

Kernel Foveated RenderingI3D '18, VR '20, TVCG '20

CollaboVR ISMAR '20

LogRectilinear

IEEE VR '21 (TVCG)

TVCG Honorable Mention

GazeChat

UIST '21

Computer Graphics

MDIF�ICCV' 21

HumanGPS�CVPR' 21

HandSight

ECCVW '14

TACCESS '15

Ad hoc UI

CHIEA '22

ProtoSound

CHI ‘22

PRIF�ECCV' 22

Computer

Vision

Rapsai

CHI ‘23

Visual Captions

ThingShare

CHI ‘23

6 of 294

Self Intro

Ruofei Du (杜若飞)

Geollery CHI '19, Web3D '19, VR '19

Social Street View Web3D '16

Best Paper Award

VideoFields

Web3D '16

SketchyScene

TOG (SIGGRAPH Asia) '19, ECCV '18

Montage4D

I3D '18

JCGT '19

DepthLab UIST '20

13K Installs

Kernel Foveated RenderingI3D '18, VR '20, TVCG '20

CollaboVR ISMAR '20

LogRectilinear

IEEE VR '21 (TVCG)

TVCG Honorable Mention

GazeChat

UIST '21

Digital World

Digital Human

MonoAvatar

HumanGPS�CVPR' 21, '23

HandSight

ECCVW '14

TACCESS '15

ProtoSound

CHI ‘22

Ad hoc UI

CHIEA '22

OmniSyn

IEEE VR '22

SlurpAR DIS '22

Visual Captions

ThingShare

CHI ‘23

Rapsai

CHI ‘23

Interaction and Communication

7 of 294

Interactive Perception & Graphics for

A Universally Accessible Metaverse

Ruofei Du

Senior Research Scientist / Manager

Google AR, San Francisco

www.ruofeidu.com

me@duruofei.com

8 of 294

Metaverse

9 of 294

How Metaverse is defined by

academia and industry?

10 of 294

Neal Stephenson, 1992.

11 of 294

Metaverse

12 of 294

13 of 294

14 of 294

15 of 294

The Future of Internet

Internet of Things

Virtual Reality

Augmented Reality

Decentralization

Blockchain

NFT

Mirrored World

Metaverse

Digital Twin

VR OS

Web 3.0

Extended Reality (XR)

Accessibility

Avatars

Co-presence

Economics

Gaming

Wearable

AI

Privacy

Security

Vision

Neural

16 of 294

How do I define Metaverse?

17 of 294

Metaverse envisioned

a persistent digital world,

where people are fully connected as

virtual representations.

18 of 294

More importantly, what research

directions shall we devote to Metaverse?

19 of 294

20 of 294

metaversea medium to make information more useful and accessible and help people to live a better physical life

21 of 294

Interactive Perception & Graphics

for a Universally Accessible Metaverse

Chapter One · Mirrored World & Real-time Rendering

Chapter Two · Computational Interaction: Algorithm & Systems

Chapter Three · Digital Human & Augmented Communication

22 of 294

Interactive Perception & Graphics for a Universally Accessible Metaverse

Chapter One · Mirrored World & Real-time Rendering

Geollery CHI '19, Web3D '19, VRW '19

Social Street View Web3D '16

Best Paper Award

Kernel Foveated RenderingI3D '18, VR '20, TVCG '20

LogRectilinear, OmniSyn

IEEE VR '21 (TVCG), VRW ‘22

TVCG Honorable Mention

23 of 294

Geollery.com & Social Street View: Reconstructing a Live Mirrored World With Geotagged Social Media

Ruofei Du, David Li, and Amitabh Varshney

{ruofei, dli7319, varshney}@umiacs.umd.edu | www.Geollery.com | ACM CHI 2019 & Web3D 2016 Best Paper Award & 2019

UMIACS

THE AUGMENTARIUM

VIRTUAL AND AUGMENTED REALITY LAB

AT THE UNIVERSITY OF MARYLAND

COMPUTER SCIENCE

UNIVERSITY OF MARYLAND, COLLEGE PARK

24 of 294

Introduction

Social Media

24

image courtesy: plannedparenthood.org

25 of 294

Introduction

Social Media + Topics

25

image courtesy: huffingtonpost.com

26 of 294

Motivation

Social Media + XR

26

27 of 294

Motivation

Social Media + XR

27

image courtesy:

instagram.com,

facebook.com,

twitter.com

28 of 294

Motivation

2D layout

28

image courtesy:

pinterest.com

29 of 294

Motivation

Immersive Mixed Reality?

29

image courtesy:

viralized.com

30 of 294

Motivation

Pros and cons of the classic

30

31 of 294

Motivation

Pros and cons of the classic

31

32 of 294

Related Work

Social Street View, Du and Varshney

Web3D 2016 Best Paper Award

32

33 of 294

34 of 294

Related Work

Social Street View, Du and Varshney

Web3D 2016 Best Paper Award

34

35 of 294

Related Work

Social Street View, Du and Varshney

Web3D 2016 Best Paper Award

35

36 of 294

Related Work

3D Visual Popularity

Bulbul and Dahyot, 2017

36

37 of 294

Related Work

Virtual Oulu, Kukka et al.

CSCW 2017

37

38 of 294

Related Work

Immersive Trip Reports

Brejcha et al. UIST 2018

38

39 of 294

Related Work

High Fidelity, Inc.

39

40 of 294

Related Work

Facebook Spaces, 2017

40

41 of 294

What's Next?

Research Question 1/3

41

What may a social media platform look like in mixed reality?

42 of 294

What's Next?

Research Question 2/3

42

What if we could allow social media sharing in a live mirrored world?

43 of 294

What's Next?

Research Question 3/3

43

What use cases can we benefit from social media platform in XR?

44 of 294

Geollery.com

A Mixed-Reality Social Media Platform

44

45 of 294

Geollery.com

A Mixed-Reality Social Media Platform

45

46 of 294

System Overview

Geollery Workflow

46

47 of 294

Geollery.com

v2: a major leap

47

48 of 294

System Overview

Geollery Workflow

48

49 of 294

System Overview

2D Map Data

49

50 of 294

System Overview

2D Map Data

50

51 of 294

System Overview

+Avatar +Trees +Clouds

51

52 of 294

System Overview

+Avatar +Trees +Clouds +Night

52

53 of 294

System Overview

Street View Panoramas

53

54 of 294

System Overview

Street View Panoramas

54

55 of 294

System Overview

Street View Panoramas

55

56 of 294

System Overview

Geollery Workflow

56

All data we used is publicly and widely available on the Internet.

57 of 294

Rendering Pipeline

Close-view Rendering

57

58 of 294

Rendering Pipeline

Initial spherical geometries

58

59 of 294

Rendering Pipeline

Depth correction

59

60 of 294

Rendering Pipeline

Intersection removal

60

61 of 294

Rendering Pipeline

Texturing individual geometry

61

62 of 294

Rendering Pipeline

Texturing with alpha blending

62

63 of 294

Rendering Pipeline

Rendering result in the fine detail

63

64 of 294

Rendering Pipeline

Rendering result in the fine detail

64

65 of 294

Rendering Pipeline

Rendering result in the fine detail

65

66 of 294

User Study

Social Street View vs. Geollery

66

67 of 294

User Study

Quantitative Evaluation

67

68 of 294

User Study

Quantitative Evaluation

68

69 of 294

69

I would like to use it for the food in different restaurants. I am always hesitating of different restaurants. It will be very easy to see all restaurants with street views. In Yelp, I can only see one restaurant at a time.

P6 / F

70 of 294

70

[I will use it for] exploring new places. If I am going on vacation somewhere, I could immerse myself into the location. If there are avatars around that area, I could ask questions.

P1 / M

71 of 294

71

I think it (Geollery) will be useful for families. I just taught my grandpa how to use Facetime last week and it would great if I could teleport to their house and meet with them, then we could chat and share photos with our avatars.

P2 / F

72 of 294

72

if there is a way to unify the interaction between them, there will be more realistic buildings [and] you could have more roof structures. Terrains will be interesting to add on.

P18 / M

73 of 294

Rendering Pipeline

Experimental Features

73

74 of 294

Landing Impact

Demos at ACM CHI 2019

74

75 of 294

Landing Impact

Demos at ACM CHI 2019

75

76 of 294

Landing Impact

Demos at ACM CHI 2019

76

77 of 294

Instant Panoramic Texture Mapping with Semantic Object Matching for Large-Scale Urban Scene Reproduction

TVCG 2021, Jinwoo Park, Ik-beom Jeon, Student Members, Sung-eui Yoon, and Woontack Woo

78 of 294

Instant Panoramic Texture Mapping with Semantic Object Matching for Large-Scale Urban Scene Reproduction

TVCG 2021, Jinwoo Park, Ik-beom Jeon, Student Members, Sung-eui Yoon, and Woontack Woo

A more applicable method for constructing walk-through experiences in urban streets was employed by Geollery [16], which adopted an efficient transformation of a dense spherical mesh to construct a local proxy geometry based on the depth maps from Google Street View

79 of 294

Freeman et al. ACM PHCI 2022

He et al. ISMAR 2020

Park et al. Virtual Reality 2022

Yeom et al. IEEE VR 2021

80 of 294

What's Next?

81 of 294

Video Fields: Fusing Multiple Surveillance Videos into a Dynamic Virtual Environment

82 of 294

image courtesy: university of maryland, college park

Introduction

Surveillance Videos

83 of 294

Architecture

Video Fields Flowchart

84 of 294

85 of 294

OmniSyn: Intermediate View Synthesis Between Wide-Baseline Panoramas

David Li, Yinda Zhang, Christian Häne, Danhang Tang, Amitabh Varshney, and Ruofei Du, VR 2022

86 of 294

OmniSyn: Intermediate View Synthesis Between Wide-Baseline Panoramas

David Li, Yinda Zhang, Christian Häne, Danhang Tang, Amitabh Varshney, and Ruofei Du, VR 2022

87 of 294

input 1

input 2

input 3

88 of 294

89 of 294

How can we further accelerate the real-time rendering procedure?

90 of 294

90

Kernel Foveated Rendering

Xiaoxu Meng, Ruofei Du, Matthias Zwicker and Amitabh Varshney

Augmentarium | UMIACS

University of Maryland, College Park

ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games 2018

91 of 294

91

Original Frame

Buffer

Screen

Sample Map

Introduction

Related Work

Our Approach

User Study

Experiments

Conclusion

92 of 294

92

 

 

 

 

Introduction

Related Work

Our Approach

User Study

Experiments

Conclusion

 

Kernel Log-polar Mapping

 

93 of 294

93

Introduction

Related Work

Our Approach

User Study

Experiments

Conclusion

94 of 294

Can we further accelerate it?

95 of 294

Eye-dominance-guided�Foveated Rendering

Xiaoxu Meng, Ruofei Du, and Amitabh Varshney

IEEE Transactions on Visualization and Computer Graphics (TVCG)

96 of 294

fovea

fovea

more foveation for the non-dominant eye

97 of 294

What if we apply to 3D volume data formats?

98 of 294

3D-Kernel Foveated Rendering for Light Fields

Xiaoxu Meng, Ruofei Du, Joseph JaJa, and Amitabh Varshney

IEEE Transactions on Visualization and Computer Graphics (TVCG), 2020

UMIACS

99 of 294

100 of 294

How about 360 video streaming?

101 of 294

A Log-Rectilinear Transformation for Foveated 360-Degree Video Streaming

David Li, Ruofei Du, Adharsh Babu, Camelia Brumar, Amitabh Varshney

University of Maryland, College Park Google Research

UMIACS

TVCG Honorable Mentions Award

102 of 294

103 of 294

104 of 294

With recent advances of neural networks,

how can we further

compress existing graphics?

105 of 294

Sandwiched Image Compression

Wrapping Neural Networks Around a Standard Codec

Increasing the Resolution and Dynamic Range of Standard Codecs

Onur Guleryuz, Philip Chou, Hugues Hoppe, Danhang Tang, �Ruofei Du, Philip Davidson, and Sean Fanello

2021 IEEE International Conference on Image Processing (ICIP)�2022 Picture Coding Symposium (PCS) Best Paper Finalist

106 of 294

107 of 294

108 of 294

109 of 294

What about compressing 3D volumes?

110 of 294

What are levels of details?

111 of 294

112 of 294

Multiresolution Deep Implicit Functions for 3D Shape Representation

Zhang Chen, Yinda Zhang, Kyle Genova, Thomas Funkhouse, Sean Fanello, Sofien Bouaziz, Christian Häne, Ruofei Du, Cem Keskin, and Danhang Tang

2021 IEEE/CVF International Conference on Computer Vision (ICCV)

113 of 294

114 of 294

115 of 294

Interactive Perception & Graphics for a Universally Accessible Metaverse

Chapter Two · Computational Interaction: Algorithm & Systems

Ad hoc UI

CHI EA ‘22

DepthLab

UIST '20

13K Installs & deployed in Tiktok, Snap, Teamviewer etc.

SlurpARDIS ‘22

Rapsai

CHI ‘23

116 of 294

DepthLab: Real-time 3D Interaction with Depth Maps for Mobile Augmented Reality

Ruofei Du, Eric Turner, Maksym Dzitsiuk, Luca Prasso, Ivo Duarte,

Jason Dourgarian, Joao Afonso, Jose Pascoal, Josh Gladstone, Nuno Cruces,

Shahram Izadi, Adarsh Kowdle, Konstantine Tsotsos, David Kim

Google | ACM UIST 2020

117 of 294

Introduction

Mobile Augmented Reality

118 of 294

Introduction

Google's ARCore

119 of 294

Introduction

Google's ARCore

120 of 294

Introduction

Mobile Augmented Reality

121 of 294

Introduction

Motivation

Is the current generation of object placement sufficient for realistic AR experiences?

122 of 294

Introduction

Depth Lab

Not always!

123 of 294

Introduction

Depth Lab

Virtual content looks like it’s “pasted on the screen” rather than “in the world”!

124 of 294

Introduction

Motivation

125 of 294

126 of 294

Introduction

Depth Lab

How can we bring these advanced

features to mobile AR experiences WITHOUT relying on dedicated sensors or the need for computationally expensive surface reconstruction?

127 of 294

Introduction

Depth Map

128 of 294

Introduction

Depth Lab

Google

Pixel 2, Pixel 2 XL, Pixel 3, Pixel 3 XL, Pixel 3a, Pixel 3a XL, Pixel 4, Pixel 4 XL

Huawei

Honor 10, Honor V20, Mate 20 Lite, Mate 20, Mate 20 X, Nova 3, Nova 4, P20, P30, P30 Pro

LG

G8X ThinQ, V35 ThinQ, V50S ThinQ, V60 ThinQ 5G

OnePlus

OnePlus 6, OnePlus 6T, OnePlus 7, OnePlus 7 Pro, OnePlus 7 Pro 5G, OnePlus 7T, OnePlus 7T Pro

Oppo

Reno Ace

Samsung

Galaxy A80, Galaxy Note8, Galaxy Note9, Galaxy Note10, Galaxy Note10 5G, Galaxy Note10+, Galaxy Note10+ 5G, Galaxy S8, Galaxy S8+, Galaxy S9, Galaxy S9+, Galaxy S10e, Galaxy S10, Galaxy S10+, Galaxy S10 5G, Galaxy S20, Galaxy S20+ 5G, Galaxy S20 Ultra 5G

Sony

Xperia XZ2, Xperia XZ2 Compact, Xperia XZ2 Premium, Xperia XZ3

Xiaomi

Pocophone F1

129 of 294

Introduction

Depth Lab

Is there more to realism than occlusion?

130 of 294

Introduction

Depth Lab

Surface interaction?

131 of 294

Introduction

Depth Lab

Realistic Physics?

132 of 294

Introduction

Depth Lab

Path Planning?

133 of 294

134 of 294

Introduction

Depth Lab

135 of 294

Related Work

Valentin et al.

136 of 294

Depth Maps

137 of 294

Depth �from Motion

Depth From a Single Camera

138 of 294

Best Practices

Depth From a Single Camera

Use depth-certified ARCore devices

Minimal movement in the scene

Encourage users to move the device

Depth from 0 to 8 meters

Best accuracy 0.5 to 5 meters

139 of 294

Enhancing Depth

Optimized to give you the best depth

Depth from Motion is fused with state-of-the-art Machine Learning

Depth leverages specialized hardware like a Time-of-Flight sensor when available

140 of 294

Introduction

Depth Lab

141 of 294

Introduction

Depth Lab

142 of 294

Introduction

Depth Generation

143 of 294

Introduction

Depth Lab

144 of 294

Related Work

Valentin et al.

145 of 294

Introduction

Depth Lab

146 of 294

Introduction

Depth Lab

Up to 8 meters, with

the best within 0.5m to 5m

147 of 294

Motivation

Gap from raw depth to applications

148 of 294

Introduction

Depth Lab

ARCore

Depth API

DepthLab

Mobile AR developers

149 of 294

Design Process

3 Brainstorming Sessions

3 brainstorming sessions

18 participants

39 aggregated ideas

150 of 294

Design Process

3 Brainstorming Sessions

151 of 294

System

Architecture overview

152 of 294

Data Structure

Depth Array

2D array (160x120 and above) of 16-bit integers

153 of 294

Data Structure

Depth Mesh

154 of 294

Data Structure

Depth Texture

155 of 294

System

Architecture

156 of 294

Localized Depth

Coordinate System Conversion

157 of 294

Localized Depth

Normal Estimation

158 of 294

Localized Depth

Normal Estimation

159 of 294

Localized Depth

Normal Estimation

160 of 294

Localized Depth

Avatar Path Planning

161 of 294

Localized Depth

Rain and Snow

162 of 294

Surface Depth

Use Cases

163 of 294

Surface Depth

Physics collider

Physics with depth mesh.

164 of 294

Surface Depth

Texture decals

Texture decals with depth mesh.

165 of 294

Surface Depth

3D Photo

Projection mapping with depth mesh.

166 of 294

Dense Depth

Depth Texture - Antialiasing

167 of 294

Dense Depth

Real-time relighting

θ

N

L

168 of 294

Dense Depth

Why normal map does not work?

169 of 294

Dense Depth

Real-time relighting

170 of 294

Dense Depth

Real-time relighting

171 of 294

Dense Depth

Real-time relighting

go/realtime-relighting, go/relit

172 of 294

Dense Depth

Wide-aperture effect

173 of 294

Dense Depth

Occlusion-based rendering

174 of 294

175 of 294

Experiments

DepthLab minimum viable application

176 of 294

Experiments

General Profiling of MVP

177 of 294

Experiments

Relighting

178 of 294

Experiments

Aperture effects

179 of 294

Impact

Deployment with partners

180 of 294

Impact

Deployment with partners

181 of 294

Impact

Deployment with partners

182 of 294

AR Realism

In TikTok

183 of 294

AR Realism

Built into Lens Studio for Snapchat Lenses

Kevaid

Saving Chelon

Quixotical�The Seed: World of Anthrotopia

Snap�Dancing Hotdog

184 of 294

Camera Image

3D Point Cloud

Provides a more detailed representation of the geometry of the objects in the scene.

Raw Depth API

New depth capabilities

185 of 294

Camera Image

Raw Depth Image

Depth Image

Confidence Image

New depth capabilities

Raw Depth API

Provides a more detailed representation of the geometry of the objects in the scene.

186 of 294

Try it yourself!

TeamViewer�LifeAR App

ARCore�Depth Lab App

Depth �Hit Test

New depth capabilities

187 of 294

ARCore�Depth Lab App

Depth API�Codelab

Raw Depth API�Codelab

188 of 294

Limitations

Design space of dynamic depth

Dynamic Depth? HoloDesk, HyperDepth, Digits, Holoportation for mobile AR?

189 of 294

Envision

Design space of dynamic depth

190 of 294

GitHub

Please feel free to fork!

191 of 294

Play Store

Try it yourself!

192 of 294

Impact

Significant Media Coverage

193 of 294

Impact

Significant Media Coverage

194 of 294

More Links

Significant Media Coverage

195 of 294

DepthLab: Real-time 3D Interaction with Depth Maps for Mobile Augmented Reality

Ruofei Du, Eric Turner, Maksym Dzitsiuk, Luca Prasso, Ivo Duarte,

Jason Dourgarian, Joao Afonso, Jose Pascoal, Josh Gladstone, Nuno Cruces,

Shahram Izadi, Adarsh Kowdle, Konstantine Tsotsos, David Kim

Google | ACM UIST 2020

196 of 294

Thank you!

DepthLab | UIST 2020

197 of 294

Demo

DepthLab | UIST 2020

198 of 294

DepthLab: Real-time 3D Interaction with Depth Maps for Mobile Augmented Reality

Ruofei Du, Eric Turner, Maksym Dzitsiuk, Luca Prasso, Ivo Duarte,

Jason Dourgarian, Joao Afonso, Jose Pascoal, Josh Gladstone, Nuno Cruces,

Shahram Izadi, Adarsh Kowdle, Konstantine Tsotsos, David Kim

Google | ACM UIST 2020

199 of 294

After exploring interaction with the environment,

how shall we interact with everyday object?

200 of 294

Ad hoc UI: On-the-fly Transformation of Everyday Objects

into Tangible 6DOF Interfaces for AR

Ruofei Du, Alex Olwal, Mathieu Le Goc, Shengzhi Wu, Danhang Tang,

Yinda Zhang, Jun Zhang, David Joseph Tan, Federico Tombari, David Kim

Google | CHI 2022 Interactivity

201 of 294

202 of 294

Applications

203 of 294

Can we learn from the history to

interact with everyday object?

204 of 294

“Slurp” Revisited: Using Software Reconstruction to Reflect on Spatial Interactivity and Locative Media

Shengzhi Wu, Daragh Byrne, Ruofei Du, and Molly Steenson

ACM DIS 2022

205 of 294

206 of 294

RetroSphere: Self-Contained Passive 3D Controller Tracking for Augmented Reality

Ananta Narayanan Balaji, Clayton Kimber, David Li, Shengzhi Wu, Ruofei Du, David Kim

ACM Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT) 2022

207 of 294

Ananta Narayanan Balaji, Clayton Kimber, David Li, Shengzhi Wu, Ruofei Du, David Kim

ACM Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT) 2022

208 of 294

209 of 294

With recent advances of on-device ML models,

how can we accelerate the prototyping efforts?

210 of 294

What if we can build applications as if building Legos?

211 of 294

rapsai

Accelerating Machine Learning Prototyping of Multimedia Applications through Visual Programming

Ruofei Du, Na Li, Jing Jin, Michelle Carney, Scott Miles, Maria Kleiner, Xiuxiu Yuan, Yinda Zhang,

Anuva Kulkarni, Xingyu "Bruce" Liu, Ahmed Sabie, Sergio Escolano, Abhishek Kar,

Ping Yu, Ram Iyengar, Adarsh Kowdle, and Alex Olwal

212 of 294

213 of 294

Interactive Graphics for a Universally Accessible Metaverse

Chapter Three · Digital Human & Augmented Communication

HumanGPS CVPR ‘21

Montage4D

I3D '18

JCGT '19

GazeChat & CollaboVRUIST ‘21 & ISMAR ‘20

Visual Captions & ThingShareCHI ‘23

214 of 294

What is Avatar?

215 of 294

216 of 294

Avatar

History & Definition

Avatar is a term used in Hinduism for a material manifestation of a deity: “descent of a deity from a heaven”

217 of 294

Avatar

History & Definition

In computing, an avatar is a graphical representation of a user or the user's character or persona.

218 of 294

Avatar

Taxonomy

What is the oldest avatar in computer history?

219 of 294

Avatar

History & Definition

220 of 294

Avatar

History & Definition

Guo, Kaiwen, Peter Lincoln, Philip Davidson, Jay Busch, Xueming Yu, Matt Whalen, Geoff Harvey et al. "The relightables: Volumetric performance capture of humans with realistic relighting." ACM Transactions on Graphics (ToG) 38, no. 6 (2019): 1-19.

221 of 294

Dating back to real-time digital human / avatars…

222 of 294

222

223 of 294

224 of 294

225 of 294

226 of 294

227 of 294

228 of 294

229 of 294

230 of 294

231 of 294

232 of 294

233 of 294

234 of 294

235 of 294

236 of 294

What is the state-of-the-art since then?

237 of 294

Related Work

Fusing Multiple Dynamic Videos

238 of 294

ACM Trans. Graph., Vol. 40, No. 4, Article 1. SIGGRAPH 2021

239 of 294

Photorealistic Characters

The Relightables

Kaiwen Guo, Peter Lincoln, Philip Davidson, Jay Busch, Xueming Yu, Matt Whalen, Geoff Harvey, Sergio Orts-Escolano, Rohit Pandey, Jason Dourgarian, Danhang Tang, Anastasia Tkach, Adarsh Kowdle, Emily Cooper, Mingsong Dou, Sean Fanello, Graham Fyffe, Christoph Rhemann, Jonathan Taylor, Paul Debevec, and Shahram Izadi. 2019. The Relightables: Volumetric Performance Capture of Humans With Realistic Relighting. ACM Transactions on Graphics, pp. . DOI: https://doi.org/10.1145/3355089.3356571

240 of 294

ACM Trans. Graph., Vol. 40, No. 4, Article 1. SIGGRAPH 2021

241 of 294

Photorealistic Characters

Rocketbox

Mar Gonzalez-Franco, Eyal Ofek, Ye Pan, Angus Antley, Anthony Steed, Bernhard Spanlang, Antonella Maselli, Domna Banakou, Nuria Pelechano, Sergio Orts Escolano, Veronica Orvahlo, Laura Trutoiu, Markus Wojcik, Maria V. Sanchez-Vives, Jeremy Bailenson, Mel Slater, and Jaron Lanier "The Rocketbox library and the utility of freely available rigged avatars." Frontiers in Virtual Reality DOI: 10.3389/frvir.2020.561558

242 of 294

Photorealistic Characters

From phone scan

Chen Cao, Tomas Simon, Jin Kyu Kim, Gabe Schwartz, Michael Zollhoefer, Shun-Suke Saito, Stephen Lombardi, Shih-En Wei, Danielle Belko, Shoou-I Yu, Yaser Sheikh, and Jason Saragih. 2022. Authentic Volumetric Avatars From a Phone Scan. ACM Transactions on Graphics, pp. . DOI: https://doi.org/10.1145/3528223.3530143

243 of 294

How can we build dynamic dense correspondence

within the same subject and

among different subjects?

244 of 294

245 of 294

246 of 294

247 of 294

248 of 294

249 of 294

How can we leverage real-time Avatars today?

250 of 294

GazeChat

Enhancing Virtual Conferences With

Gaze-Aware 3D Photos

Zhenyi He, Keru Wang, Brandon Yushan Feng, Ruofei Du, Ken Perlin

New York University� University of Maryland, College Park � Google

251 of 294

252 of 294

Introduction

VR headset & video streaming

252

253 of 294

Related Work

Gaze-2 (2003)

253

254 of 294

Related Work

MultiView (2005)

254

255 of 294

Related Work

MMSpace (2016)

255

256 of 294

Our Work

GazeChat (UIST 2021)

256

257 of 294

Gaze Awareness

Definition

257

Gaze awareness, defined here as knowing what someone is looking at.

258 of 294

Gaze Awareness

Definition

258

gaze correction

gaze redirection

raw input image

GazeChat

259 of 294

Gaze Correction

Definition

259

260 of 294

Gaze Rediction

Definition

260

eye contact

who is looking at whom

261 of 294

Pipeline

System

261

262 of 294

Eye Tracking

WebGazer..js

262

263 of 294

Neural Rendering

Eye movement

263

264 of 294

Neural Rendering

Eye movement

264

265 of 294

3D Photo Rendering

3D photos

265

266 of 294

3D Photo Rendering

3D photos

266

267 of 294

Layouts

UI

267

268 of 294

Networking

WebRTC

268

269 of 294

How can we work in XR as stylized avatars?

270 of 294

Zhenyi He* Ruofei Du Ken Perlin*

*Future Reality Lab, New York University Google LLC

CollaboVR: A Reconfigurable Framework for

Creative Collaboration in Virtual Reality

271 of 294

272 of 294

How can we further augment communication,

in videoconferencing, AR, and XR in future?

273 of 294

Visual Captions

Augmenting Verbal Communication

With On-the-Fly Visuals

Xingyu "Bruce" Liu, Vladimir Kirilyuk, Xiuxiu Yuan, Peggy Chi,

Xiang "Anthony" Chen, Alex Olwal, and Ruofei Du

274 of 294

275 of 294

ThingShare

Ad-Hoc Digital Copies of Physical Objects for Sharing Things in Video Meetings

Erzhen Hu, Jens Emil Grønbæk, Wen Ying, Ruofei Du, and Seongkook Heo

276 of 294

277 of 294

How can AI benefit a broader inclusive community?

278 of 294

ProtoSound: A Personalized and Scalable Sound Recognition System for Deaf and Hard-of-Hearing Users

ACM CHI 2012 · Dhruv Jain, Khoa Nguyen, Steven Goodman, Rachel Grossman-Kahn, Hung Ngo, Aditya Kusupati, Ruofei Du, Alex Olwal, Leah Findlater, and Jon Froehlich

279 of 294

280 of 294

How can AI + Metaverse improve our life?

281 of 294

SketchyScene: Richly-Annotated Scene Sketches

Changqing Zou, Qian Yu, Ruofei Du, Haoran Mo, Yi-Zhe Song, Tao Xiang, Chengying Gao, Baoquan Chen, and Hao Zhang (ECCV 2022)

282 of 294

Language-based Colorization of Scene Sketches

Changqing Zou, Haoran Mo, Chengying Gao, Ruofei Du, and Hongbo Fu (ACM Transaction on Graphics, SIGGRAPH Asia 2019)

283 of 294

284 of 294

285 of 294

286 of 294

Future Directions

The Ultimate XR Platform

286

287 of 294

Future Directions

The Ultimate XR Platform

287

288 of 294

Future Directions

Fuses Past Events

288

289 of 294

Future Directions

With the present

289

290 of 294

Future Directions

And look into the future

290

291 of 294

Future Directions

Change the way we communicate in 3D and consume the information

291

292 of 294

Future Directions

Consume the information throughout the world

292

293 of 294

Interactive Perception & Graphics for a

Universally Accessible Metaverse

Ruofei Du

Senior Research Scientist / Manager

Google AR

www.ruofeidu.com

me@duruofei.com

Thank you!

294 of 294

CollaboVR: A Reconfigurable Framework for

Creative Collaboration in Virtual Reality