Augmented Reality Object Identification via Social Keyword Extraction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current augmented reality applications have a low identification success rate for random objects in environments without markers, limiting their applicability and user experience.

Innovation Solution

A method that collects user-uploaded images and label information, extracts keywords from social networking comments, and generates augmented reality patterns and content based on image features and keywords, allowing for automatic identification of random objects without markers.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If traditional AR identification methods are used for special images (landmarks, books, barcodes), then identification accuracy is high, but the types of identifiable objects are limited and application scenarios are restricted

Engineering Contradiction:
Improvetypes of identifiable objectsVSAvoididentification success rate
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent creates virtual copies of real-world objects by generating 3D models from multiple 2D images captured from different angles. These 3D models serve as reusable templates that can be matched against new images, enabling the system to identify not only pre-defined special images but also random objects in the environment, thereby expanding object types while maintaining identification accuracy through robust 3D-2D matching algorithms

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system performs preliminary actions by pre-capturing multiple images of objects from different viewpoints and generating their 3D models in advance. This preparatory work creates a library of object representations that can be quickly matched during AR identification, allowing the system to handle both special images and random objects efficiently without requiring real-time complex processing

Inventive Principle:
Principle #10Preliminary action

2Reliability

If AR software processes data from sensors to identify objects in photos, then identification can be achieved for special images, but identification success rate for random objects without markers remains low

Engineering Contradiction:
Improveidentification success rate for random objectsVSAvoidcomplexity of identification system
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces 3D models as an intermediary representation between 2D images and object identification. Instead of directly matching 2D images which varies in perspective and lighting, the system converts images to 3D models and performs matching in 3D space, which is more robust to viewpoint changes. This intermediary representation significantly improves reliability for identifying random objects while keeping the processing pipeline manageable through automated algorithms

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system transitions from 2D image matching to 3D model matching by constructing three-dimensional representations from multiple two-dimensional images. This dimensional elevation allows the system to capture objects from arbitrary viewpoints and perform robust matching in 3D space, dramatically improving identification success rate for random objects without requiring complex real-time 2D pattern recognition

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Extent of automation

If user images and social networking comments are collected and processed, then augmented reality patterns and content can be automatically generated for random objects, but data processing complexity increases

Engineering Contradiction:
Improveautomatic generation of AR contentVSAvoidcomplexity of data processing system
Core Design Contradiction:
Extent of automationVSDevice complexity

Solution Approach 1:

The system implements self-service by automatically generating AR patterns and content through algorithms that process user-uploaded images and social networking comments. The automated pipeline extracts object features, creates 3D models, generates identification patterns, and retrieves relevant content without manual intervention, achieving high extent of automation while managing complexity through modular processing stages and standardized data formats

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP2851811B1Method and device for achieving augmented reality application
Publication Date: 2019.03.13 HUAWEI DEVICE CO LTD
  • EP2851811B1 patent drawingFigure 1
  • EP2851811B1 patent drawingFigure 2
  • EP2851811B1 patent drawingFigure 3

AI summary

The present invention discloses a method for implementing an augmented reality application. The method including: collecting an image uploaded by a user and label information of the image; releasing the image and the label information to a social networking contact of the user according to the user's social graph and interest graph on the Internet, and obtaining the social networking contact's comment information about the image; extracting, from the comment information, a keyword of which occurrence frequency is higher than a first threshold; adding the image to an image album according to the label information of the image and the keyword; and generating an augmented reality pattern and augmented reality content about a describing object of the image according to image features of all images in the image album and the keyword. The present invention further discloses a device for implementing an augmented reality application. Embodiments of the present invention can solve a problem of identifying a random object in an environment without a marker in an augmented reality application.