Cell Image Segmentation via Spatial Frequency Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In images of cells cultured in a medium, distinguishing cell images from background and debris becomes challenging as culture progresses, especially when using simple processing methods like binarization or edge detection.

Innovation Solution

An image processing method that segments images based on local spatial frequency components, using two-dimensional discrete cosine transform (DCT) to differentiate between cell, background, and debris by representing blocks with low frequency components for brightness and high frequency components for texture, allowing precise extraction of cell areas.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If simple processing methods like binarization or edge detection are used, then the processing speed is fast and easy to implement, but the accuracy of distinguishing cell images from debris deteriorates

Engineering Contradiction:
Improveaccuracy of distinguishing cell images from debrisVSAvoidcomplexity of processing method
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent divides the image into multiple blocks and processes each block independently to extract spatial frequency components. This segmentation allows detailed analysis of local image characteristics without processing the entire image at once, improving accuracy while managing computational complexity through distributed processing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transforms image data from the spatial domain to the frequency domain using two-dimensional discrete cosine transform. This dimensional transformation enables distinction between cell images and debris based on spatial frequency characteristics, providing a new perspective for accurate segmentation.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If traditional imaging methods are used, then the cell area can be specified with existing techniques, but the ability to observe cell changes over time without modifying the cell deteriorates

Engineering Contradiction:
Improveability to observe cell changes over timeVSAvoidease of cell modification process
Core Design Contradiction:
Adaptability or versatilityVSEase of manufacture

Solution Approach 1:

The patent replaces physical cell modification methods (chemical staining, fluorescence tagging) with an image processing-based approach. By using spatial frequency analysis on grayscale images, the system can identify cell boundaries and track cell changes over time without altering the cells, enabling non-invasive long-term observation.

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

3Measurement precision

If the image is divided into blocks and spatial frequency components are extracted, then the accuracy of cell area extraction is improved, but the processing time increases

Engineering Contradiction:
Improveprecision of cell area extractionVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the necessary spatial frequency components (low frequency and high frequency components) from each block rather than processing all frequency components. This selective extraction maintains sufficient accuracy for cell area identification while reducing computational burden and processing time.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3770853B1Image processing method, computer program, and recording medium
Publication Date: 2024.08.28 SCREEN HOLDINGS CO LTD
  • EP3770853B1 patent drawingFigure 1
  • EP3770853B1 patent drawingFigure 2
  • EP3770853B1 patent drawingFigure 3A~3C

AI summary

To segment properly a cell area and another area from an original image including a cell and a background, an image processing method according to the present invention includes obtaining an original image including a cultured cell image with a background image (Step S101), dividing the original image into blocks each composed of a predetermined number of pixels and obtaining a spatial frequency component of an image in each block for each block (Step S104-S106), and classifying each block as the one belonging to a cell cluster corresponding to the cell or the one belonging to other than the cell cluster in a two-dimensional feature amount space composed of a first feature amount which is a total of intensities of low frequency components having a frequency equal to or lower than a predetermined frequency and a second feature amount which is a total of intensities of high frequency components having a higher frequency than the low frequency component (Step S107-S108), and segmenting the original image into an area occupied by the blocks classified as the cell cluster and another area (Step S109).