Automated Image Annotation via Machine-Readable ID Linking

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Manual annotation of images for training machine learning models in computer vision is time-consuming and costly, often lacking the accuracy required for high reliability, especially when automatic annotation methods fall short in providing sufficient accuracy.

Innovation Solution

A method that utilizes a computer vision system with an image capturing device and a reader system to automatically annotate images by linking machine-readable identifiers with subject IDs, employing machine learning techniques such as deep learning to detect and localize subjects, and train models on annotated data using frameworks like Faster-RCNN or pre-trained DNNs to improve annotation accuracy and efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual annotation is used to annotate images for training machine learning models, then the accuracy of labeled data is improved, but the time and cost required for annotation increases significantly

Engineering Contradiction:

Improveaccuracy of labeled dataVSAvoidtime required for annotation

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary actions by using automatic annotation methods to pre-annotate images before human review. This preliminary annotation provides a head start, reducing the total time needed for achieving high-quality labeled data while maintaining accuracy through subsequent human verification steps.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements feedback loops where automatically annotated images are reviewed and corrected by human annotators, and the corrected annotations are fed back into the system to improve future automatic annotation accuracy. This iterative feedback process gradually reduces the time required for manual correction while maintaining high accuracy standards.

Inventive Principle:

Principle #23Feedback

2Measurement precision

If manual annotation is used to annotate images for training machine learning models, then the accuracy of labeled data is improved, but the cost of annotation increases significantly

Engineering Contradiction:

Improveaccuracy of labeled dataVSAvoidcost of annotation

Core Design Contradiction:

Measurement precisionVSLoss of energy

Solution Approach 1:

The system performs preliminary automatic annotation to create initial labeled data sets, reducing the amount of expensive manual annotation work required. This preliminary action significantly lowers costs while maintaining acceptable accuracy levels that are further improved through selective human review.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system uses feedback from human annotation reviews to continuously improve the automatic annotation algorithms, reducing the proportion of expensive manual work needed over time. As the automatic system learns from feedback, the cost ratio shifts favorably toward automation while maintaining high accuracy.

Inventive Principle:

Principle #23Feedback

3Productivity

If automatic annotation methods are used to annotate images, then the time and cost of annotation are reduced, but the accuracy of labeled data decreases

Engineering Contradiction:

Improveannotation efficiencyVSAvoidaccuracy of labeled data

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system segments the annotation process into multiple stages: automatic annotation for initial labeling, human review for correction and validation, and iterative improvement cycles. This segmentation allows the system to leverage the speed of automatic methods while incorporating human accuracy where most needed, achieving both high productivity and high accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system creates a composite annotation approach combining automatic annotation algorithms with human expert review, similar to composite materials combining different substances for optimal properties. This hybrid approach achieves annotation speeds close to fully automatic methods while maintaining accuracy接近 to fully manual annotation through strategic human intervention.

Inventive Principle:

Principle #40Composite materials

Data Source

PatentUS20240212385A1Method and system for automatically annotating and identifying a living being or an object with an identifier providing a subject identification

Publication Date: 2024.06.27 KEPLER VISION TECH BV

AI summary

The invention relates to a method for training a machine learning model to identify a subject having at least one machine readable identifier providing a subject ID, said method comprising:providing a computer vision system with an image capturing system comprising at least one image capturing device, and a reader system comprising at least one reader for reading said at least one machine readable identifier;defining said machine learning model in said computer vision system;capturing a first image using said image capturing system, said first image showing said subject;reading said subject ID using said reader system when capturing said first image, and linking said subject ID with said first image, said linking providing said first image with a linked subject ID, providing a first annotated image;capturing at least one further image showing said subject, linking said linked subject ID to said at least one further image providing at least one further annotated image, andsubjecting said first annotated image and said at least one further annotated image to said machine learning model for training said machine learning model.