Photorealistic Avatar Animation for Gaze-Aware Virtual Conferencing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In virtual conferences, participants' gaze directions and speech are not accurately conveyed, leading to difficulties in determining who is looking at whom and issues with user experience when cameras are turned off due to bandwidth or privacy concerns.

Innovation Solution

The use of machine learning models to generate synthesized images based on participant photos, combined with gaze and speech data, to animate photorealistic avatars in virtual conferencing systems, reducing the need for live video streaming and enhancing user experience.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If live video streaming is used to convey participant appearance and gaze, then accuracy of gaze and speech representation is improved, but bandwidth consumption increases and privacy concerns arise

Engineering Contradiction:
Improvegaze direction accuracyVSAvoidbandwidth consumption
Core Design Contradiction:
Measurement precisionVSLoss of energy

Solution Approach 1:

The patent creates photorealistic avatar copies of participants that replicate their appearance, gaze directions, and speech characteristics. These avatars are synthesized from static photos and audio recordings, allowing the system to convey participant presence without streaming live video, thus reducing bandwidth consumption while maintaining representation accuracy

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system performs preliminary actions by capturing static photos and audio recordings of participants before the conference, then uses these pre-captured materials to synthesize avatars. This preliminary capture eliminates the need for continuous live video streaming during the conference, reducing bandwidth consumption while preserving the ability to accurately represent participant gaze and speech

Inventive Principle:
Principle #10Preliminary action

2Reliability

If cameras are turned off to protect privacy and reduce bandwidth usage, then privacy is protected and bandwidth is reduced, but user experience deteriorates due to loss of visual presence

Engineering Contradiction:
Improveprivacy protectionVSAvoiduser experience
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The patent creates photorealistic avatar copies of participants that replicate their appearance, gaze directions, and speech characteristics. These avatars are synthesized from static photos and audio recordings, allowing the system to convey participant presence without streaming live video, thus reducing bandwidth consumption while maintaining representation accuracy

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The photorealistic avatars serve as intermediaries between the actual participants and other conference attendees. The avatars convey participant presence, gaze, and speech characteristics without requiring participants to activate cameras, thus mediating the interaction while protecting privacy and reducing bandwidth usage

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If multiple synthesized images are generated for different gaze directions, then accuracy of gaze representation is improved, but computational complexity increases

Engineering Contradiction:
Improvegaze direction representationVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by generating multiple synthesized images with different gaze directions in advance, before the conference begins. These pre-synthesized images are stored and then selected based on the participant's actual gaze direction during the conference, reducing real-time computational complexity while maintaining gaze representation accuracy

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically selects and blends appropriate synthesized images based on the participant's current gaze direction. By using interpolation and blending techniques, the system can generate intermediate gaze directions from discrete synthesized images, achieving continuous gaze representation without synthesizing images for every possible angle, thus balancing accuracy with computational efficiency

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4385203B1Avatar animation in virtual conferencing
Publication Date: 2026.02.11 GOOGLE LLC
  • EP4385203B1 patent drawingFigure 1A
  • EP4385203B1 patent drawingFigure 1B
  • EP4385203B1 patent drawingFigure 1C~1D

AI summary

According to a general aspect, a method can include receiving a photo of a virtual conference participant, and a depth map based on the photo, and generating a plurality of synthesized images based on the photo. The plurality of synthesized images can have respective simulated gaze directions of the virtual conference participant. The method can also include receiving, during a virtual conference, an indication of a current gaze direction of the virtual conference participant. The method can further include animating, in a display of the virtual conference, an avatar corresponding with the virtual conference participant. The avatar can be based on the photo. Animating the avatar can be based on the photo, the depth map and at least one synthesized image of the plurality of synthesized images, the at least one synthesized image corresponding with the current gaze direction.