3D Participant Representation Using Partial Realtime Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Capturing and encoding 3D imaging data for virtual meetings results in high network bandwidth demands and latency issues, particularly due to the large amounts of data required for real-time 3D streams, which negatively impact the conferencing experience.
Innovation Solution
Combining partial real-time 3D data with a non-real-time 3D model to create a combined 3D representation, reducing bandwidth and latency requirements by transmitting only partial real-time data, which is then combined with a pre-generated model to provide a complete 3D representation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complete real-time 3D data is transmitted for each participant, then the 3D virtual meeting quality is improved, but the network bandwidth consumption increases significantly
Solution Approach 1:
The patent segments the 3D data transmission into two parts: a pre-generated non-realtime 3D model (capturing static features like face shape, hair, clothing) and partial real-time 3D data (capturing only dynamic features like facial expressions and head movements). This segmentation allows the system to transmit only the necessary dynamic portions in real-time, significantly reducing bandwidth consumption while maintaining overall 3D representation quality.
Solution Approach 2:
The patent performs preliminary action by generating and storing a complete non-realtime 3D model of each participant before the virtual meeting begins. This pre-generated model contains all static anatomical and appearance features. During the meeting, only the dynamic real-time portions need to be transmitted and combined with the pre-existing model, reducing the real-time data transmission requirement.
2Measurement precision
If complete real-time 3D data is captured and transmitted, then the realism of participant representation is improved, but the network latency and jitter increase
Solution Approach 1:
The patent separates the 3D data into static components (pre-generated model) and dynamic components (real-time capture). By transmitting only the dynamic portions during the meeting, the data volume is reduced, which directly decreases network transmission time and latency, while still maintaining realism through the combination with the detailed pre-generated model.
Solution Approach 2:
By pre-generating and storing the complete static 3D model before the meeting, the system eliminates the need to transmit this large data set in real-time. This preliminary action reduces the real-time data transmission burden, thereby reducing network latency and jitter while preserving the ability to render realistic representations.
3Adaptability or versatility
If 3D imaging is implemented for all participants, then the immersive experience is improved, but the resource usage (bandwidth, processing) increases
Solution Approach 1:
The patent segments the resource-intensive 3D imaging process into a resource-heavy pre-processing phase (generating non-realtime models before meetings) and a resource-light real-time phase (capturing only dynamic portions). This allows the system to provide immersive 3D experiences during meetings while significantly reducing the network resource usage during the actual meeting sessions.
Solution Approach 2:
By performing the computationally intensive 3D model generation in advance (before the meeting), the patent shifts the resource consumption to a time when network resources are less constrained. During the meeting, only lightweight real-time data transmission is required, reducing the immediate network resource usage while maintaining the immersive 3D experience.
Data Source
AI summary
A method for providing a three-dimensional (3D) representation of a transmitting participant in a virtual meeting is provided. The method is performed in a representation provider and comprises obtaining a non-realtime 3D model of at least part of a person, obtaining partial realtime 3D data of the transmitting participant of the virtual meeting, and combining the non-realtime 3D model with the partial realtime 3D data, resulting in a combined 3D representation of the transmitting participant.