Automated Annotation via Reference Image Point Mapping

Solution Overview

Problem

Existing data annotation techniques for machine learning rely heavily on human operators, requiring hundreds of hours to annotate large volumes of input images, which is inefficient and costly.

Innovation Solution

A method that involves receiving a reference image with annotations and mapping points from the reference image to corresponding points in input images, allowing the annotations to be projected onto the input images, thereby automating the annotation process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If human operators manually annotate each input image, then annotation accuracy can be maintained, but the time and cost required increases significantly

Engineering Contradiction:

Improveannotation accuracyVSAvoidannotation time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary action by manually annotating a single reference image, then uses automated point mapping and projection to transfer these annotations to multiple input images. This preliminary annotation of one reference image replaces the need for manual annotation of each individual input image, dramatically reducing annotation time while maintaining consistency through the mapping process.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system creates copies of annotations from the reference image and projects them onto multiple input images through point mapping. Instead of creating original annotations for each image, the system copies the annotation structure and adapts it to different images based on corresponding point relationships, enabling rapid annotation generation.

Inventive Principle:

Principle #26Copying

2Reliability

If human operators manually annotate each input image, then annotation quality can be ensured, but the cost and resource requirements increase

Engineering Contradiction:

Improveannotation qualityVSAvoidannotation throughput

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system performs preliminary action by manually annotating a single reference image, then uses automated point mapping and projection to transfer these annotations to multiple input images. This preliminary annotation of one reference image replaces the need for manual annotation of each individual input image, dramatically reducing annotation time while maintaining consistency through the mapping process.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system creates copies of annotations from the reference image and projects them onto multiple input images through point mapping. Instead of creating original annotations for each image, the system copies the annotation structure and adapts it to different images based on corresponding point relationships, enabling rapid annotation generation.

Inventive Principle:

Principle #26Copying

3Productivity

If automated point mapping is used to project annotations, then annotation efficiency increases, but the complexity of the annotation system increases

Engineering Contradiction:

Improveannotation efficiencyVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system introduces an intermediary point mapping mechanism that bridges the reference image and input images. By establishing corresponding points between images and using these points as intermediaries for annotation projection, the system automates the annotation process while managing complexity through a structured intermediate representation rather than direct image-to-image annotation transfer.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS12293563B2Automated data annotation for computer vision applications

Publication Date: 2025.05.06 SYNAPTICS INC

AI summary

This disclosure provides methods, devices, and systems for training machine learning models. The present implementations more specifically relate to techniques for automating the annotation of data for training machine learning models. In some aspects, a machine learning system may receive a reference image depicting an object of interest with one or more annotations and also may receive one or more input images depicting the object of interest at various distances, angles, or locations but without annotations. The machine learning system maps a set of points in the reference image to a respective set of points in each input image so that the annotations from the reference image are projected onto the object of interest in each input image. The machine learning system may further train a machine learning model to produce inferences about the object of interest based on the annotated input images.