Time-Lapse Image Positional Correction via Feature-Weighted Correlation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image-processing methods for time-lapse images face challenges in accurately correcting positional shifts between intermittently captured images, especially when foreign objects or varying cell characteristics affect the assumption of random movement, leading to inaccuracies in positional correction.

Innovation Solution

An image-processing method that calculates positional correction values using a correlation function weighted according to image features of objects in the field of view, allowing for precise alignment between images, even with non-target objects or varying cell dynamics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If positional correction is performed by obtaining a correlation of an image as a whole between respective images, then the complexity and cost of the apparatus is reduced by avoiding hardware alignment systems, but the accuracy of positional correction is compromised when foreign objects or non-target articles are present in the field of view

Engineering Contradiction:
Improveapparatus complexityVSAvoidpositional correction accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent segments the image processing by identifying and separating target objects (cells) from non-target objects (foreign objects, bubbles) in the field of view. The correlation calculation is then performed only on regions containing target objects, excluding regions with non-target objects. This segmentation approach maintains low apparatus complexity while improving positional correction accuracy by preventing foreign objects from interfering with the correlation-based alignment process.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If hardware alignment systems or costly low thermal expansion materials are used to minimize relative positional shift, then positional correction accuracy is improved, but the apparatus complexity and cost increase significantly

Engineering Contradiction:
Improvepositional correction accuracyVSAvoidapparatus complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces mechanical alignment systems (hardware alignment optical systems) and material-based solutions (Super Invar, ceramic) with a software-based image processing approach. By using correlation calculations on segmented image regions containing only target objects, the system achieves accurate positional correction without requiring complex mechanical alignment mechanisms or costly low thermal expansion materials, thus reducing apparatus complexity while maintaining measurement precision.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Ease of operation

If the assumption of random cell movement is made for positional correction, then the processing method is simplified, but the accuracy deteriorates when foreign objects or articles not forming part of the observed object are present

Engineering Contradiction:
Improveprocessing simplicityVSAvoidpositional correction accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent extracts and removes non-target objects (foreign objects, bubbles, articles not forming part of the observed object) from the image regions used for correlation calculation. By taking out these interfering elements from the analysis, the system maintains the simplicity of the correlation-based processing method while improving accuracy, as the positional correction is now based only on the movement patterns of actual target objects (cells) rather than being contaminated by foreign objects.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP2234061B1Image processing method for time lapse image, image processing program, and image processing device
Publication Date: 2019.11.27 NIKON CORP
  • EP2234061B1 patent drawingFigure 1
  • EP2234061B1 patent drawingFigure 2
  • EP2234061B1 patent drawingFigure 3

AI summary

An image-processing method for processing a time lapse image comprises: a step (S110) for acquiring a first image obtained by capturing an image of an object located within a visual field using an imaging device and a second image obtained by capturing an image of the object located within the visual field using the imaging device after a predetermined time period has elapsed; a step (S180) for calculating a positional correction value for a positional shift between the first image and the second image, by using a correlation function weighted according to an image feature value of each of the objects included in the fist image and the second image; and a step (S190) for performing, based on the calculated positional correction value, a positional correction between the first image and the second image. The time lapse image is generated using the first image and the second image on which the positional correction has been performed.