3D Avatar Capture Using Progressive Multi-View Body Scanning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing representations of users, such as avatars, are not realistic depictions of the user operating an electronic device due to limitations in capturing the user's full appearance, especially when sensors have a limited field of view or when body parts are occluded.
Innovation Solution
Progressively capture 3D appearance data over time using multiple sensors with limited fields of view, combining image data from different viewpoints to build a composite 3D representation that updates as the user interacts with the device, incorporating machine learning to refine the model and include temporal characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If multiple sensors with limited fields of view are used to capture user appearance data, then the completeness of the 3D avatar representation is improved over time, but the complexity of the system increases
Solution Approach 1:
The system divides the capture task into segments by using multiple sensors with limited fields of view, each capturing different portions of the user's body at different times. The 3D avatar is built incrementally by combining these segmented views, resolving the contradiction between using simple single-sensor systems and complex multi-sensor systems.
Solution Approach 2:
The system performs preliminary capture actions over time, accumulating appearance data for different body parts as they come into view. This progressive capture approach allows the system to build a complete avatar representation without requiring all sensors to be active simultaneously, reducing immediate system complexity while improving long-term data completeness.
2Measurement precision
If sensors capture only currently visible body parts, then the accuracy of the avatar representation is improved, but the time required to build a complete avatar increases
Solution Approach 1:
The system maintains continuous capture operations as the user moves, continuously updating the 3D avatar representation with newly captured body parts. This continuous process resolves the contradiction by ensuring that every moment of sensor operation contributes to building the complete avatar, maximizing data collection efficiency without sacrificing accuracy.
Solution Approach 2:
The system dynamically adapts the avatar representation based on currently captured data, updating the model as new body parts become visible. This dynamic approach allows the system to maintain high accuracy for captured portions while progressively expanding coverage over time, balancing accuracy requirements with time constraints.
Data Source
AI summary
Various implementations disclosed herein include devices, systems, and methods that progressively capture data representing an actual appearance of a user for creating a 3D avatar of the user. Image sensors at a user's electronic device may capture images (or other sensor data) of different portions of a user's body over time. Images and other sensor data that is captured initially or at a given time may not represent all of the user's body. Thus, in some implementations, the progressively captured data provides representations of additional portions of the user's 3D avatar over time.