Keypoint Unwarping for Machine Vision Feature Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing machine vision algorithms face challenges in accurately matching features across images captured with different opto-electronic systems due to image distortion, leading to unreliable feature matching and machine vision performance.

Innovation Solution

The implementation of a keypoint unwarping method that transforms keypoints based on an image deformation model, allowing for reliable feature matching and machine vision processing across images with varying distortions by unwarping only the keypoints, reducing computational resources and avoiding full-image unwarping issues like blurring.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If full-image unwarping is applied to correct image distortion, then feature matching reliability is improved, but processing time and computational resources increase significantly

Engineering Contradiction:

Improvefeature matching reliabilityVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent segments the image processing task by applying unwarping only to specific keypoint regions rather than the entire image. The system identifies keypoint locations, extracts local image patches around these keypoints, and applies unwarping transformation only to these small regions. This segmentation approach maintains feature matching reliability while dramatically reducing processing time and computational resources compared to full-image unwarping.

Inventive Principle:

Principle #1Segmentation

2Reliability

If full-image unwarping is applied to correct image distortion, then feature matching reliability is improved, but computational resources and processing complexity increase

Engineering Contradiction:

Improvefeature matching reliabilityVSAvoidprocessing complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent divides the image into small local patches centered at keypoint locations and processes only these segments. By segmenting the processing task, the system avoids the complexity of full-image unwarping while maintaining matching reliability. The segmentation is combined with selective processing based on keypoint importance and local distortion characteristics.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different processing quality levels to different regions of the image. High-quality unwarping is applied only to regions containing keypoints that are critical for matching, while other regions receive minimal or no processing. This local quality approach optimizes the balance between reliability and complexity by concentrating computational resources where they are most needed.

Inventive Principle:

Principle #3Local quality

3Reliability

If full-image unwarping is applied to correct image distortion, then feature matching across different opto-electronic systems is improved, but image quality deteriorates due to blurring

Engineering Contradiction:

Improvefeature matching reliabilityVSAvoidimage quality

Core Design Contradiction:

ReliabilityVSManufacturing precision

Solution Approach 1:

The patent applies unwarping transformation only to small local patches around keypoints rather than the entire image. This localized approach prevents the cumulative blurring effects that occur with full-image unwarping, as each small patch undergoes minimal transformation. The segmentation ensures that only the critical keypoint regions are modified, preserving overall image quality while enabling cross-system feature matching.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent maintains high image quality by applying high-quality unwarping only locally at keypoint regions where it is most needed for matching accuracy. The rest of the image retains its original quality without undergoing degradation from unwarping operations. This local quality preservation is achieved by limiting the spatial extent of unwarping transformations.

Inventive Principle:

Principle #3Local quality

4Productivity

If traditional feature matching algorithms are used without unwarping, then processing speed is maintained, but matching accuracy deteriorates due to image distortion

Engineering Contradiction:

Improveprocessing speedVSAvoidmatching accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent performs preliminary unwarping correction on small keypoint patches before the feature matching process. By applying the unwarping transformation in advance to the critical keypoint regions, the system ensures that matching algorithms operate on corrected data, improving accuracy without sacrificing overall processing speed. This preliminary action is localized to keypoint regions rather than the entire image.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11475238B2Keypoint unwarping for machine vision applications

Publication Date: 2022.10.18 STMICROELECTRONICS SRL

AI summary

An image processing system has one or more memories and image processing circuitry coupled to the one or more memories. The image processing circuitry, in operation, compares a first image to feature data in a comparison image space using a matching model. The comparing includes: unwarping keypoints in keypoint data of the first image; and comparing the unwarped keypoints and descriptor data associated with the first image to the feature data of the comparison image. The image processing circuitry determines whether the first image matches the comparison image based on the comparing.