Dynamic Reference Image Generation for Multi-User AR Pose Estimation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional multi-user augmented reality systems require a previously acquired common reference, limiting their usage in unknown environments, and existing technologies struggle to accurately estimate poses for multiple users without prior knowledge of the environment.

Innovation Solution

A system that generates a reference image on the fly by capturing images of a planar object from different viewpoints, determining the pose of a first mobile platform, and warping the image to produce a front view, which is then used by other mobile platforms to estimate their pose relative to the same planar object, allowing for multi-user augmented reality applications without prior environmental knowledge.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a previously acquired common reference is used for tracking, then pose estimation accuracy is improved, but the system cannot operate in unknown environments

Engineering Contradiction:
Improvepose estimation accuracyVSAvoidenvironment compatibility
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary actions by having the first mobile platform capture images and determine pose information before other platforms join. This pre-established reference framework allows subsequent platforms to estimate poses accurately without requiring pre-acquired environmental references, resolving the contradiction between measurement precision and environment compatibility.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary mechanism where the first mobile platform acts as a temporary reference frame for other platforms. Instead of requiring all platforms to reference a pre-existing common model, each platform estimates its pose relative to the first platform's captured images, enabling operation in unknown environments while maintaining pose estimation accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of operation

If motion sensors are used to generate front view images, then pose estimation is enabled, but additional components are required and sensor errors are introduced

Engineering Contradiction:
Improvepose estimation capabilityVSAvoidcomponent requirements
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The patent replaces the mechanical sensor-based system with a vision-based system. Instead of relying on accelerometers and gyroscopes to generate front view images, the system uses image processing and geometric transformations on captured images to achieve pose estimation, eliminating the need for additional motion sensors and their associated errors.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system creates a virtual copy of the front view by warping captured images through geometric transformation rather than using physical sensors. This digital copying approach replaces the need for motion sensors while maintaining the ability to generate accurate front view representations for pose estimation.

Inventive Principle:
Principle #26Copying

3Measurement precision

If mean patches from limited viewpoints are used, then pose estimation is achieved, but the method is limited to vertical or horizontal planar surfaces

Engineering Contradiction:
Improvepose estimationVSAvoidsurface orientation coverage
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent introduces dynamic adaptability by allowing the warping transformation to adjust to any planar surface orientation. Instead of using fixed mean patches for limited viewpoints, the system dynamically transforms captured images based on the actual pose and orientation of the planar surface, enabling accurate pose estimation for surfaces in any orientation.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes the parameters of the transformation model to accommodate different surface orientations. By using general geometric warping transformations rather than fixed mean patches, the system can adapt to any planar surface orientation while maintaining pose estimation accuracy, thus resolving the limitation to vertical or horizontal surfaces.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2614487B1Online reference generation and tracking for multi-user augmented reality
Publication Date: 2021.01.06 QUALCOMM INC
  • EP2614487B1 patent drawingFigure 1~2
  • EP2614487B1 patent drawingFigure 3~4
  • EP2614487B1 patent drawingFigure 5

AI summary

A multi-user augmented reality (AR) system operates without a previously acquired common reference by generating a reference image on the fly. The reference image is produced by capturing at least two images of a planar object and using the images to determine a pose (position and orientation) of a first mobile platform with respect to the planar object. Based on the orientation of the mobile platform, an image of the planar object, which may be one of the initial images or a subsequently captured image, is warped to produce the reference image of a front view of the planar object. The reference image may be produced by the mobile platform or by, e.g., a server. Other mobile platforms may determine their pose with respect to the planar object using the reference image to perform a multi-user augmented reality application.