AR Depth Image Blending for Stable 3D Object Occlusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current augmented reality (AR) systems suffer from inaccurate and unstable depth data, leading to inconsistent rendering of 3D objects relative to real-world objects, resulting in a less than desirable user experience.

Innovation Solution

Storing depth data associated with real-world objects in a buffer and using this data to accurately render 3D objects by blending and replacing missing depth information with valid data, utilizing surface elements (surfels) to enhance depth and position processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If current AR systems use depth data from depth cameras, then real-time 3D reconstruction is achieved, but depth data becomes inaccurate and unstable causing incorrect rendering of 3D objects

Engineering Contradiction:
Improvereal-time 3D reconstruction speedVSAvoiddepth data accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The system performs preliminary actions by capturing and storing depth images and surfels before the actual rendering occurs. Historical depth data is accumulated in advance to create a comprehensive depth buffer that can be referenced during rendering, ensuring accurate depth information is available even when current depth camera data is unreliable.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary approach by using surfels (surface elements) as a mediator between the depth camera and the rendering system. Surfels serve as intermediate representations that bridge the gap between raw depth data and final rendering, allowing for more accurate and stable depth information to be derived and used in the rendering process.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Speed

If depth data is processed in real-time, then rendering responsiveness is improved, but depth information becomes unstable and inaccurate

Engineering Contradiction:
Improverendering responsivenessVSAvoiddepth data stability
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The system performs preliminary actions by capturing and storing depth images and surfels before the actual rendering occurs. Historical depth data is accumulated in advance to create a comprehensive depth buffer that can be referenced during rendering, ensuring accurate depth information is available even when current depth camera data is unreliable.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback mechanisms where the system continuously monitors and refines depth information based on historical data and rendering results. The depth buffer is updated and refined through feedback loops that compare current depth data with historical surfel data, adjusting and correcting inaccuracies in real-time to improve both responsiveness and stability.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If multiple depth images are stored and processed, then depth accuracy is improved, but system complexity and data processing requirements increase

Engineering Contradiction:
Improvedepth information accuracyVSAvoiddata processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the depth data into discrete surfels (surface elements) that represent specific portions of the real-world scene. This segmentation allows the system to process and store depth information in manageable units, reducing the complexity of handling large volumes of depth data while maintaining high accuracy through the accumulated surfel representations.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4154222B1Method and apparatus for combining an augmented reality object in a real-world image
Publication Date: 2026.01.28 GOOGLE LLC
  • EP4154222B1 patent drawingFigure 1A
  • EP4154222B1 patent drawingFigure 1B
  • EP4154222B1 patent drawingFigure 1C

AI summary

A method including receiving a first depth image associated with a first frame at a first time of an augmented reality (AR) application, the first depth image representing at least a first portion of a real-world space storing the first depth image receiving a second depth image associated with a second frame at a second time, after the first time, of the AR application, the second depth image representing at least a second portion of the real-world space generating a real-world image by blending, at least, the stored first depth image with the second depth image receiving a rendered AR object combining the AR object in the real-world image and displaying the real-world image combined with the AR object.