AR Virtual Object Localization via Neural Network Probability Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality (AR) localization methods face challenges in achieving high accuracy and commercialization due to the need for expensive sensors and difficulties in dynamic environments, especially when projecting virtual objects onto real-world images with many moving objects.
Innovation Solution
A method that generates a probability map using a trained neural network to determine the directional characteristic of objects, projects virtual objects onto map data based on localization information, and adjusts this information through visual alignment between input and projected images, allowing for accurate placement and display of virtual objects in AR services.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor-based localization method uses high accuracy sensors, then localization precision is improved, but device cost and complexity increase
Solution Approach 1:
The patent replaces sensor-based mechanical localization systems with a vision-based computational approach. Instead of using expensive physical sensors like GPS and IMU, the system uses camera images processed through neural networks to achieve localization, thereby reducing hardware complexity while maintaining precision
Solution Approach 2:
The patent creates a virtual copy of the real-world environment through map data and projects virtual objects onto it. This digital twin approach allows localization to be achieved through image processing and visual alignment rather than expensive physical sensors
2Device complexity
If vision-based localization method is used, then device cost is reduced, but reliability decreases in dynamic environments
Solution Approach 1:
The patent performs preliminary actions by pre-processing images to generate probability maps and directional characteristics before localization. The neural network pre-computes object directions and probabilities, creating a robust foundation for reliable localization even in dynamic environments
Solution Approach 2:
The patent implements feedback mechanisms through visual alignment between projected virtual images and actual camera images. The system continuously compares and adjusts localization based on the alignment degree, improving reliability by correcting errors in real-time
3Adaptability or versatility
If virtual objects are projected onto map data, then AR service functionality is improved, but visual alignment accuracy deteriorates due to localization errors
Solution Approach 1:
The patent makes the localization system dynamic by allowing continuous adjustment of localization information based on visual alignment feedback. The system adapts to environmental changes and corrects positioning errors in real-time, maintaining accurate visual alignment while preserving AR functionality
Solution Approach 2:
The patent changes localization parameters (position, orientation) based on visual alignment measurements. By adjusting these parameters iteratively to maximize alignment between virtual and real images, the system maintains precision while enabling versatile AR services
Data Source
AI summary
Disclosed is a localization method and apparatus that may acquire localization information of a device, generate a first image that includes a directional characteristic corresponding to an object included in an input image, generate a second image in which the object is projected based on the localization information, to map data corresponding to a location of the object, and adjust the localization information based on visual alignment between the first image and the second image.