Object Identification Using Digital Signatures for Edge Devices

Solution Overview

Problem

Existing Augmented Reality applications face challenges in efficiently identifying objects of interest within images captured by moving image capture devices due to the processing-intensive nature of 3D mapping and localization, which becomes cumbersome and time-consuming, especially in real-time scenarios.

Innovation Solution

The method utilizes digital signatures and context associated with reference images to identify objects of interest within images generated by a relocatable image capture device, employing a query processing module trained with neural networks to recognize these objects efficiently, allowing for real-time detection and augmentation of images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If 3D mapping and localization methods are used to identify objects of interest, then object identification accuracy is improved, but processing time and computational complexity increase significantly

Engineering Contradiction:

Improveobject identification accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary actions by pre-computing and storing digital signatures of objects of interest in a database before actual image processing. These digital signatures are generated in advance from reference images and stored for quick retrieval during real-time object identification, eliminating the need for complex 3D mapping at runtime

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent creates simplified copies of objects in the form of digital signatures that capture essential visual characteristics without full 3D geometric detail. These digital signature copies are stored in a database and used for rapid comparison against query images, providing accurate object identification with minimal processing overhead

Inventive Principle:

Principle #26Copying

Solution Approach 3:

The system extracts only the essential visual features needed for object identification into compact digital signatures, separating these key characteristics from the full 3D point cloud data. This extraction allows the system to retain object identification accuracy while eliminating unnecessary computational complexity

Inventive Principle:

Principle #2Taking out (Extraction)

2Reliability

If exhaustive back end processing is used for 3D localization, then object identification reliability is improved, but processing efficiency deteriorates

Engineering Contradiction:

Improveobject identification reliabilityVSAvoidprocessing efficiency

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent replaces the mechanical 3D mapping and localization processing system with an information-based digital signature matching system. Instead of performing computationally intensive 3D geometric transformations and point cloud comparisons, the system uses digital signature extraction and database lookup, dramatically improving processing efficiency while maintaining reliability

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes the processing parameters from full 3D point cloud data to compact digital signature representations. This parameter transformation reduces the data dimensionality and computational complexity from exhaustive 3D localization to efficient digital signature matching, enabling real-time processing without sacrificing object identification reliability

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If full 3D point cloud data is processed for object identification, then identification accuracy is improved, but computational load and resource requirements increase

Engineering Contradiction:

Improveobject identification accuracyVSAvoidcomputational load

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system extracts only the essential visual features needed for object identification into compact digital signatures, separating these key characteristics from the full 3D point cloud data. This extraction allows the system to retain object identification accuracy while eliminating unnecessary computational complexity

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent creates simplified copies of objects in the form of digital signatures that capture essential visual characteristics without full 3D geometric detail. These digital signature copies are stored in a database and used for rapid comparison against query images, providing accurate object identification with minimal processing overhead

Inventive Principle:

Principle #26Copying

Data Source

PatentUS12073615B2Method, apparatus, and computer program product for identifying objects of interest within an image captured by a relocatable image capture device

Publication Date: 2024.08.27 HERE GLOBAL BV

AI summary

A method, apparatus and computer program product identify objects of interest within images. In a method, the method receives one or more images generated by an image capture device that is configured to be relocated. The method also receives information regarding an estimated location of the image capture device. Based upon a representation of a respective image of the one or more images and respective representations of a plurality of reference images, the method identifies one or more reference images that are associated with the estimated location. The method also identifies an object of interest within the respective image generated by the image capture device based upon respective digital signatures of one or more objects of interest and at least some context associated with the one or more objects of interest that are depicted by the one or more reference images that have been identified.