KR19990039172A - Analysis of Cilia Movement of Cells by Image Processing - Google Patents

Abstract

The present invention relates to a method for analyzing ciliary movement of a cell, and in particular, a cell by image processing for analyzing ciliary movement using a digitized ciliated image instead of using a direct analog signal caused by the ciliary movement observed under a microscope. Ciliary motion analysis method of the present invention.

The present invention extracts the brightness signal of the A / D converted microscope image, obtains the motion frequency of the cilia using the FFT, and color-codes the spatial distribution of the dominant peak frequencies to represent the distribution map of the cilia motion frequency. It is not only effective as a clinical and research tool because it is easy to analyze the cell's kinetic frequency, it is simple to use, and it is accurate compared to the conventional methods, and it is also possible to study the cilia kinetic frequency of other organic cells similar to the cilia of the nose. Applicable to the field

Description

Analysis of Cilia Movement of Cells by Image Processing

The present invention relates to a method for analyzing ciliary movement of a cell, and in particular, a cell by image processing for analyzing ciliary movement using a digitized ciliated image instead of using a direct analog signal caused by the ciliary movement observed under a microscope. Ciliary motion analysis method of the present invention.

The ciliated epithelial cells remove the foreign substances that enter the respiratory tract, including the nasal cavity, through ciliary movement. At this time, the ciliary beat frequency (hereinafter referred to as CBF) may vary depending on the species of the animal or within the species.

In addition, autonomic drugs used in the clinic are mainly used for heart and respiratory diseases, but the effect on the nasal mucosal cilia is not well known. Clinically some sympathetic excitatory drugs can reduce mucosal ciliary movements, and thus abuse of these drugs can be the pathogenesis of the upper airway disease. Therefore, it is important to quantitatively measure the effects of these drugs on mucosal cilia and to present objective data on safe concentrations of drugs and appropriate exposure times.

For this reason, several methods have been tried so far to quantitatively measure ciliary frequency.

In vivo, cilia movement measurement studies were conducted by Dalhamn (1956) using a surgical microscope to determine the number of cilia movements using light reflected from mucus waves. Hakansson & Toremalm (1972) ) Measured the number of mucus waves according to ciliary movement in animals as ciliary frequency.

On the other hand, various in vitro experiments were also attempted for accurate quantitative measurement.Ballenger (1960) measured cilia using cell rotation by cilia. This is a measuring method using a stroboscope, which uses the principle that the cilia show the least movement when the frequency of the stroboscope is synchronized with the frequency of the cilia movement. However, this method provides a single frequency for several ciliated cells, so it is not possible to distinguish the frequency change of one cell over time or the frequency difference between neighboring individual cells, and to synchronize the cilia movement with the stroboscope. There is a possibility that different operators can derive different results, and the number of ciliary movements can be obtained only when the rotating cells are recorded.

In recent 20 years, many methods using photoelectric signals have been used to measure the movement of the cilia. This method uses the principle that when the amount of light transmitted through regular movement of the cilia occurs, the light is detected by the optical element and the optical signal is converted into an electrical signal. The spatial resolution that can be analyzed by this method is It is equivalent to about 10 cells. Thus the frequency measured is the average of the cells. In addition, a method of measuring scattered light using a laser can be limited to 1 to 2 cells (Lee & Verdugo, 1976).

Kennedy and Duckett (1981) introduced fast Fourier transform analysis (FFT analysis) to transform the amount of transmitted light, corresponding to the number of cilia in the mixed frequencies. To get the frequency. However, this method does not reflect ciliary movements of single cells, but reflects ciliary movements of cells visible in the microscope field. Therefore, it is difficult to think of the objective data as measured in each cell with different ciliary frequency.

The methods described above had the disadvantage of not being able to analyze many cells at the same time, measuring cilia in a single cell, and not knowing how large the cilia's movement was. This problem was slightly solved with a video recorder, which allows video to be played back so that many cells can be recorded and analyzed at the same time, or magnified through a camera, allowing detailed observation of cilia in each cell. Rautiainen et al. (1992) observed cilia at 400 and 1000 times, and then recorded the cilia using high-speed imaging and replayed them slowly to obtain cilia by counting the number of cilia. Observed the cilia directly, not only the number of times, but how strong they move in one move.

However, by measuring the cilia movement with the naked eye directly, the movement of the cilia with a small width can be overlooked, confused with the surrounding cilia, and the bias of the observer can enter. Therefore, a more active and quantitative analysis method was needed to supplement these points.

The present invention for solving the above problems is objective and quantitative to observe the cilia movement in a single cell, image processing and analyzing the number of cilia movements using a computer to exclude factors affecting other ciliary movements It is an object of the present invention to provide a method for analyzing cilia motility of cells by phosphorus image processing.

1 is a block diagram of an analysis system used in the present invention

Figure 2 is a flow chart showing a process of analyzing the cilia movement of cells according to the present invention

Figure 3 is a result of analyzing the characteristics of the ciliary movement of nasal respiratory epithelial cells according to an embodiment of the present invention

4 is a diagram showing an image of ciliated cells collected using the analysis system of FIG.

5 and 6 are CBFDM for the imaging of the ciliated cells of FIG.

7A and 7B show results of windowing a specific area of CBFDM.

8A and 8B show CBFDM showing the magnitude and phase of the peak frequency dominant in the CBF regional distribution for the particular region of FIG. 7A.

9 is a diagram analyzing the change of CBF according to phenylephrine concentration and exposure time

Explanation of symbols on the main parts of the drawings

100: microscope 200: CCD camera

300: VCR 400: Computer

410: image processing board 420: analysis program

In order to achieve the above object, the present invention digitizes and collects a microscope image for analysis on a computer, and then divides the collected image into zones, extracts data, analyzes the cilia motion frequency from the extracted data, and visualizes it. It is characterized by outputting.

1 is a block diagram showing the configuration of an analysis system used in the method for analyzing cilia movement of cells by image processing of the present invention.

An analysis system using an image processing technique widely used for medical purposes of FIG. 1 processes a microscope image of a cell, performs an analysis to obtain desired information from the processed image using a computer, and outputs the result. A microscope (100) for observing; A Charged Coupled Device (CCD) camera (200) mounted to the microscope (200) to convert an optical signal of the microscope (100) into an NTSC (National Television System Committee) analog video signal; A VCR (Video Cassette Recorder: 300) for playing back recorded video; Image processing board 410 for converting the analog image signal of the CCD camera 200 and the VCR 300 into a digital signal for analysis by the computer 400 and the image information collected through the image processing board 410 And a computer 400 having an analysis program 420 for analyzing the number of ciliary movements and the size of ciliary movements.

The microscope 100 uses a normal optical microscope, and the magnification of the microscope is appropriately determined by the horizontal and vertical pixel size occupied by one ciliated cell and the amount of spatial information that one pixel can represent. That is, the cilia should be magnified enough to induce a sufficient amount of light. When using an ordinary optical microscope, the magnification of about 1000 times shows a sufficient amount of light for analysis. Inverted microscopes are also suitable for viewing reactions due to drug use, and a heated chamber is required to maintain the temperature of the ciliated cells at a constant level.

Hereinafter, a method for analyzing cilia movement of cells according to an embodiment of the present invention will be described with reference to the accompanying drawings.

2 is a flow chart illustrating a process of analyzing ciliary movement of cells, and converts a direct microscope image taken by the CCD camera 200 or an indirectly recorded image of the VCR 300 into a digital signal that can be analyzed by the computer 400. Image collection process for collecting; A segmentation process of dividing the collected image into rectangular regions of a predetermined size to analyze the kinetic characteristics of one ciliated cell without circumferential influences; A data extraction process of analyzing a change amount of brightness of the divided zone; A frequency detection step of detecting the cilia motion frequency using the extracted data; And outputting a result of visualizing and outputting the extracted data and the detected cilia frequency.

This is explained in more detail.

First, the collection of images required for analysis is performed by directly collecting the microscope image using the CCD camera 200 or collecting the pre-recorded ciliary image through the VCR 300.

The direct microscope image is collected after converting the optical signal of the microscope 100 into an NTSC image signal in the CCD camera 200.

NTSC video signals consist of 30 frames per second, or 60 fields. The frequency of cilia known to date is usually around 15 Hz, so in order to prevent aliasing of motion during cilia analysis, an image acquisition speed of 30 Hz or more (30 phases per second) is required. Also, 60Hz (60 phases per second) image acquisition speed is required to analyze frequency increase of more than 15Hz. The NTSC signal thus provides enough information for analysis.

The entire field of the direct microscope image of the CCD camera 200 or the indirect recording image of the VCR 300 as described above is A / D by the image processing board 410 at 60 Hz (60 fields per second) or 30 Hz (30 frames per second). It is converted and stored in main memory through the internal bus of the computer 400 for about 2 seconds (64 frames).

In this case, the spatial size of each frame is 640 * 480 pixels, and since one pixel of each image is displayed as 8 bits, it is displayed with 256 kinds of brightness.

Once the image data is transferred to the memory, all subsequent information processing is done by the analysis program 320. That is, the analysis program 420 calculates and analyzes the necessary results from the data extracted from the collected image, and visualizes the obtained data and the analyzed results and shows them to the user.

To do this, first, the collected images are analyzed for automatic brightness change of the frame due to cilia motion in the time domain, and the fields are each 16 (4 * 4: size 160) to visualize the CBF distribution over the entire field in the spatial domain. 120 pixels) to 19200 (160 * 120 size 4 * 4 pixels). This makes it possible to identify the CBF distribution for multiple cells or for one cell.

Then, to extract the data required for analysis, two methods can be used for this purpose.

One is to use the amount of change in the total sum of the brightness of the collected image pixels, the other is to use the amount of change in the relative brightness of the other image relative to the reference initial image.

The method of using the change amount of the overall brightness is a method of analyzing the change amount after adding up the brightness of all the pixels of the separated regions of each image continuously collected. That is, the amount of brightness variation of the entire area is analyzed, and a basic data extraction equation is shown in Equation 1 below.

Where k is the number of frames and g _k (x, y) is the brightness of the pixel (x, y).

This method has a merit that the photoelectric effect method and the method in a continuous line can be precisely limited to the cell to be analyzed an area representing the amount of change in the amount of light. However, this method may not show a significant difference in the amount of change in the extracted data when the size of the region is small, and may not clearly indicate the dominant frequency component even by the FFT analysis described later.

Therefore, a method of extracting data using relative brightness is more preferable.

The following two methods can be used to derive the relative change in brightness.

One is to calculate the difference between the pixel-to-pixel brightness of the image continuously collected from the first image and square it, and then add all the values to analyze the change over time. If the cilia show periodicity, this data also has periodicity, and thus the highest value can be obtained from a value corresponding to a specific frequency in the FFT analysis described later. The data extraction equation using this method is expressed as in Equation (2).

Where k is the number of image frames and g _k (x, y) is the brightness of the pixel (x, y).

The other is accomplished by finding the sum of the brightness differences of the pixels versus the pixels, and then finding the difference by subtracting the average of the brightness differences of the zones.

That is, the value of the brightness difference for each pixel at the same position as the first frame (reference frame) collected for the divided region for data extraction is calculated. The pixel-to-pixel difference values are summed for all pixels in each divided region of m * n pixels, which are processed from the first frame (k = 0) to the last (k = 63). This yields a continuous result for every divided section of fields. In addition, the average value is calculated and subtracted from the original value (the sum of the brightness differences) in order to exclude bias in obtaining the result of a specific block.

The sum of the brightness differences and the difference obtained by subtracting the average value from the sum may be obtained by using Equations 3 and 4.

In Equations 3 and 4, i is a block number of a field, k is a contiguous image frame number, g _k (x, y) is a brightness of pixels (x, y) of frame k, and g ₀ (x, y Is the brightness of the pixels (x, y) of the reference frame. C _i (k) is the sum of the brightness differences of the blocks i and k, and d _i (k) is the sum of the brightness differences of the blocks i and k, minus the average value.

The power spectrum is prepared through FFT analysis (fast Fourier transformation analysis) from the data extracted by the above methods. Since the number of images collected for analysis is 64 frames (128 fields), the lowest frequency resolution in the power spectrum is about 0.47 (60/128) Hz.

The first highest frequency, or first harmonic, as a result of the power spectrum is defined as the fundamental CBF.

Here, the FFT analysis is a method of converting a data value (time signal) according to time change into a frequency distribution (frequency signal) according to frequency change.

If the time signal changes with one frequency (period), the FFT analysis results that only one value of the corresponding frequency has a nonzero value. Common biosignals have various distributions over several frequency bands. However, for signals that repeat periodically, such as the movement of ciliated cells, they typically exhibit peaks at specific frequency values, which can be viewed as CBFs that characterize ciliated cell movement.

The equation for FFT analysis is shown in Equation 5 below.

Where w is the frequency in the frequency dimension, N is the number of input signals in the time dimension, and k is a time variable in the time dimension.

The result of the FFT analysis is a period function with N as the period. In fact, if the conversion target is a real number rather than a complex number, the result is symmetrical with respect to half of the period. Therefore, only half of the repeated cycles provide meaningful information. That is, when images are collected and analyzed at a rate of 30 Hz (30 images / sec), only up to 15 Hz (CBF) provides meaningful information.

Using the obtained CBF, a CBF Distribution Map (hereinafter referred to as CBFDM) is constructed. The color-coded peak frequency value is displayed overlapping the area where the frequency value is calculated.

When designing the CBFDM, select an area that satisfies the condition that the size of the second peak is less than the predetermined value of the first peak (eg 80%) and ignore the rest.

That is, a color coded CBF for an area that satisfies the above condition is displayed for all the pixels in that area, and the original brightness pixel value at that position for another block.

The color coding of CBF is described in that the 128 RGB color-coded levels in the 256 color palette are used to represent frequencies from 0 to 15 Hz, and the remaining 128 brightness levels are used to display gray scale pixels. do.

The frequency 0 Hz is represented by the blue (0,0,255) palette, 7.5 Hz by green (0,255,0), and 15 Hz by red (255,0,0).

Intermediate frequencies from 0 to 7.5 Hz are linearly inserted by decreasing the blue (B value) of the palette and increasing green (G value), and frequencies from 7.5 to 15 Hz reduce the green (G value) and red ( Is inserted linearly by increasing R value).

This makes it easy to see the spatial distribution of CBF for multiple ciliated cells or for one cell in the whole field through CBFDM of various zone sizes.

In addition, the present invention includes a window function that shows the CBF characteristics for a specific region specified by the user in the process of visualizing and outputting the extracted data and the detected CBF, the window function allows the user to view the CBF characteristics of the specific cell region If you put a rectangular window on the CBFDM, the CBF distribution statistics of the regions included in the window are displayed as a histogram of the first highest frequency.

In the histogram statistics, the dominant frequency with the highest population is considered the CBF of the windowed region. Its magnitude and phase distribution can be color coded in proportion to the reference value, such as the peak frequency distribution.

3 to 9 show the results of analyzing the cilia of the nasal respiratory epithelial cells using the analysis method of the above-described embodiment.

The data extraction here is done by calculating the sum of the pixel-to-pixel brightness differences and then subtracting the average of the brightness differences in the zones. A typical example of the results extracted from one zone of the field is shown in the graph below in FIG. Is shown in. Here we see a regular pattern with distinctly recognizable cycles. The lower graph signal representing the power spectrum was obtained using the FFT shown in the upper left of FIG. 3, and the phase thereof is shown in the upper right graph.

The most dominant peak frequency for most zones is easily detected, which is considered the CBF of that zone and can be color coded to form CBFDM if the main conditions are met. This FFT analysis allows the user to examine the effects of drug concentration and exposure time on the cell's ciliary movement immediately and actively during testing.

The CBFDM of FIGS. 5 and 6 shows the spatial distribution of the dominant peak frequencies from the power spectra of various sized regions for ciliated images (640 * 480 pixels) of digitized 1000-times microscopic images of FIG. One region of FIG. 5 is 40 * 40 pixels, and one region of FIG. 6 is 10 * 10 pixels.

7A and 7B also show these statistics as the user positions the window relative to where the user recorded CBF occurrences ranging from 0 to 15 Hz to see more detailed statistics of the CBF distribution in a particular area, such as the surface of the cell. It is shown in the form of a histogram.

8A and 8B are CBFDMs showing magnitudes and phases of the dominant peak frequencies in the regional distribution of CBFs in the region shown in FIG. 7A, respectively, and FIG. 9 is phenylephrine analyzed using the analytical method of the present invention. (phenylephrine) shows the change of CBF with concentration and exposure time.

As described above, the present invention extracts the brightness signal of the A / D converted microscope image, obtains it from the power spectrum by FFT analysis, and then color-codes the spatial distribution of the dominant peak frequencies to map the ciliary motion frequency. It can be easily analyzed the frequency of movement of the ciliated cells, and is simple to use.

In addition, it is accurate as compared to the conventional method, and thus is effective as a clinical and research tool, and can be applied to fields such as the study of the cilia locomotor frequency of other organic cells similar to the cilia of the nose.

Claims (8)

In the method of analyzing the ciliary movement of cells, A method of analyzing cilia of a cell by image processing, characterized in that the cilia of the cell is digitized to analyze the cilia of the cell. The method of claim 1, A microscope (100) for observing cilia movements; A CCD camera 200 mounted on the microscope 200 to convert an optical signal of the microscope 100 into an NTSC analog image signal; A VCR 300 for playing back the recorded video; Image processing board 410 for converting the analog image signal of the CCD camera 200 and the VCR 300 into a digital signal for analysis by the computer 400 and the image information collected through the image processing board 410 Cell by image processing, characterized in that the analysis of the ciliary movement of the cell using an analysis system comprising a computer 400 having an analysis program 420 for analyzing the number of ciliary movement and the size of the ciliary movement from Method of ciliary motion in humans. The method of claim 2, An image acquisition process of converting and collecting a direct microscope image captured by the CCD camera 200 or an indirectly recorded image of the VCR 300 into a digital signal that can be analyzed by the computer 400; A segmentation process of dividing the collected image into rectangular regions of a predetermined size to analyze the kinetic characteristics of one ciliated cell without circumferential influences; A data extraction process of analyzing a change amount of brightness of the divided zone; A frequency detection step of detecting the cilia motion frequency using the extracted data; A method for analyzing cilia movement of cells by image processing, comprising: outputting a result of visualizing and outputting extracted data and detected cilia movement frequencies. The method of claim 3, The data extraction in the data extraction process sums up the brightness of all the pixels of the separated regions of each image continuously collected and changes the total amount of brightness or the difference in pixel-to-pixel brightness of the images continuously collected from the first image. After finding and squared, sum all the values to find the change in the relative brightness of the other image relative to the reference initial image, or the sum of the pixel-to-pixel brightness differences in the divided zones, and then the average of the brightness differences in the zones. Cilia movement analysis method of the cell by the image processing, characterized in that obtained by subtracting the difference. The method of claim 3, Frequency detection in the ciliary motor frequency detection process is performed by the FFT analysis method of the cell ciliary movement analysis by the image processing. The method of claim 3, The result output of the result output process is characterized in that the ciliary motion frequency detected in the divided region color-coded to display the distribution map of the cilia movement of the cell, characterized in that the display. The method of claim 6, The distribution map shows only the region where the second highest frequency of the divided region is less than or equal to the predetermined value of the first highest frequency, and the remaining regions are filled with the actual image. The method according to claim 6 or 7, When the user puts a rectangular window on the desired area in the distribution map, the cilia of the cell by the image processing comprising a window function for displaying the CBF distribution statistics of the areas included in the window by displaying a histogram of the first highest frequency. Exercise analysis method.