ML Image Injection Compensation for VR/AR Display Artifacts

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Virtual reality and augmented reality devices with transparent displays suffer from artifacts such as ghosting and blurring due to their rendering pipelines, which degrade image quality and user experience.

Innovation Solution

A machine learning model is trained to modify images before rendering, introducing inversion effects that counteract the artifacts introduced by the rendering pipeline, thereby improving image quality on transparent displays.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a rendering pipeline is used to display images on transparent displays, then the device can function as a VR/AR wearable device, but artifacts such as ghosting and blurring are introduced that degrade image quality

Engineering Contradiction:

ImproveVR/AR device functionalityVSAvoidimage quality

Core Design Contradiction:

Adaptability or versatilityVSManufacturing precision

Solution Approach 1:

The machine learning model applies preliminary anti-action by predicting and counteracting rendering pipeline artifacts before the image is displayed. The model processes the input image and generates a corrected output image that compensates for expected ghosting and blurring effects, thereby preventing quality degradation rather than merely addressing it after occurrence

Inventive Principle:

Principle #9Preliminary anti-action

Solution Approach 2:

The machine learning model serves as an intermediary between the rendering pipeline and the final display output. It processes images through a neural network that learns to invert artifact effects, acting as a mediating layer that transforms rendered images into artifact-compensated images before they reach the transparent display

Inventive Principle:

Principle #24Intermediary (Mediator)

2Manufacturing precision

If a machine learning model is used to modify images before rendering, then image quality is improved by reducing artifacts, but device complexity increases

Engineering Contradiction:

Improveimage qualityVSAvoidprocessing complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent replaces traditional mechanical or algorithmic image processing methods with a machine learning-based neural network. Instead of using conventional signal processing techniques to correct artifacts, the system employs a trained ML model that has learned artifact patterns and their corrections, substituting complex mechanical processing with intelligent computational processing

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The machine learning model changes parameters by learning optimal correction transformations during training. The model adjusts its internal weights and biases to optimize the relationship between input and output images, effectively changing the processing parameters dynamically based on learned patterns rather than using fixed correction algorithms

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS12518338B2Machine learning models for image injection that compensate for artifacts

Publication Date: 2026.01.06 GOOGLE LLC

AI summary

A method including an image is received, a modified image is generated based on the image using a trained machine learning model, a rendered image is generated based on the modified image using an image injection pipeline, and the rendered image is displayed on a display of a wearable device.