JP5487534B2 - Image analysis method and fluorescence detection apparatus - Google Patents

Description

  The present invention relates to an image analysis method for detecting the amount of fluorescence from a specimen more accurately and a fluorescence detection apparatus.

Conventionally, a high-throughput screening (hereinafter abbreviated as HTS) apparatus that collectively measures the amount of fluorescence in a well, typified by a well plate reader, has been the mainstream among apparatuses that detect minute fluorescence. The purpose of HTS is to test at high speed, and only obtain the total amount of fluorescence expressed from the target protein. When viewed at the single cell level, abnormalities such as expression at abnormal sites and excessive expression of cells Only the amount of fluorescence including proper expression was detected (see, for example, Patent Document 1).
JP 2000-14035 A

  Conventional HTS captures the total expression level or the state of expression of fluorescent proteins and does not determine whether normal fluorescent expression or abnormal fluorescent expression, so only information on normally expressed fluorescent proteins is obtained. There is a problem that says not.

In order to solve the above problems, the present invention provides:
A specimen image and a fluorescence image of the specimen are continuously acquired by a detector, and less than a threshold luminance value which is an upper limit value of a normal fluorescence expression amount in a cell unit of the specimen into which a predetermined fluorescent protein is introduced from the fluorescence image the extracts a first area of the cell units expressing the fluorescence,
A second region in which the ratio of the fluorescence expression area to the fluorescence non-expression area is within a predetermined range in each cell unit of the first region is extracted, and the texture feature amount of each cell in the second region in the sample image An image analysis method is provided that extracts a third region of the fluorescent image within a predetermined range .

In addition, the present invention continuously obtains a specimen image and a fluorescence image of a specimen with a detector, and an upper limit value of a normal fluorescence expression amount in a cell unit of the specimen into which a predetermined fluorescent protein is introduced from the fluorescence image Extracting a first region of the cell unit expressing fluorescence below a threshold luminance value of
Extract only the cells that contain the first region from the cells of the specimen image, and select the second region of the fluorescence image in which the ratio of the fluorescence expression area to the fluorescence non-expression area is within a predetermined range within each cell unit. Provided is an image analysis method characterized by extracting.

  In addition, the present invention can provide a fluorescence detection apparatus having the image analysis method.

The present invention also provides:
Detection means having a spatial resolution for continuously acquiring a specimen image and a fluorescence image of the specimen;
Threshold setting means for setting a threshold luminance value of fluorescence from the fluorescence image;
A first region extracting means for extracting a first region of a cell unit expressing fluorescence less than the threshold luminance value from the fluorescence image;
A second region extracting means for extracting a second region in which the ratio of the fluorescence expression area and the fluorescence non-expression area is within a predetermined range within each cell unit in the first region;
A third region extracting means for extracting a third region in which the texture feature amount of each cell in the second region in the sample image is within a predetermined range;
Measuring means for measuring the fluorescence expression amount from the fluorescence image corresponding to the third region ,
The threshold luminance value is an upper limit value of a normal fluorescence expression amount in the cell unit of the specimen into which a predetermined fluorescent protein is introduced.
The present invention also provides:
Detection means having a spatial resolution for continuously acquiring a specimen image and a fluorescence image of the specimen;
Threshold setting means for setting a threshold luminance value of fluorescence from the fluorescence image;
A first region extracting means for extracting a first region of a cell unit expressing fluorescence less than the threshold luminance value from the fluorescence image;
Only cells that contain the first region in the specimen image are extracted, and a second region in which the ratio of the fluorescence expression area to the fluorescence non-expression area is within a predetermined range in each cell unit in the first region is extracted. Second region extracting means for performing,
Measuring means for measuring the fluorescence expression amount from the fluorescence image corresponding to the second region,
The threshold luminance value is an upper limit value of a normal fluorescence expression amount in the cell unit of the specimen into which a predetermined fluorescent protein is introduced.

  ADVANTAGE OF THE INVENTION According to this invention, the fluorescence detection apparatus which can extract only the fluorescent protein which showed normal expression, and can acquire exact introduction efficiency can be provided. Moreover, the image analysis method used for this can be provided.

  Hereinafter, a fluorescence detection apparatus according to an embodiment of the present invention and an image analysis method used therefor will be described with reference to the drawings.

  FIG. 1 shows a schematic configuration of a fluorescence detection apparatus according to an embodiment. The fluorescence detection apparatus according to the embodiment is common in the image analysis methods of the first and second embodiments described later.

  In FIG. 1, a specimen 2 such as a well plate in which cells are placed is placed on a stage 3 that can move in a plane. A transmission illumination optical system 4 for obtaining a phase difference image of the sample 2 is disposed above the sample 2 (above the paper surface). The light emitted from the specimen 2 after being illuminated by the transmission illumination optical system 4 is collected by the objective lens 5 disposed below the specimen 2 (below the paper surface), and incident on the filter turret 6 via the phase difference filter 13. After passing through, the image is formed on the imaging device 8 such as a CCD by the imaging lens 7. At this time, fluorescent cubes 6a and the like, which will be described later, disposed on the filter turret 6 are removed from the optical path.

  The formed phase difference image of the sample 2 is picked up by the image pickup device 8, processed by the image processing circuit 10 of the control device 9, stored as a phase difference image of the sample 2 in a memory (not shown), and connected to the control device 9. Displayed on the display device 11. The control device 9 is a personal computer (PC) or the like, and the display device 11 is a PC monitor or the like.

  After acquiring the phase difference image of the sample 2, a shutter or the like (not shown) is opened to switch the illumination light to the sample 2 to the excitation illumination optical system 12 for fluorescence observation. At this time, the illumination light from the transmission illumination optical system 4 is controlled by the control device 9 so that the illumination light is blocked by a built-in shutter (not shown) and the sample 2 is not irradiated. The phase difference filter 13 is also moved out of the optical path.

  The excitation illumination optical system 12 includes a plurality of light sources 12 a, 12 b, and 12 c having different wavelengths, and light emitted from the light sources 12 a to 12 c is reflected and transmitted by the plurality of dichroic mirrors 12 d and half mirrors 12 e to the filter turret 6. Incident. In addition, light sources, such as a laser and LED, are used for the light sources 12a-12c.

  At the time of fluorescence observation, the filter turret 6 includes a dichroic mirror 6d that reflects a predetermined wavelength and a fluorescent cube 6a that includes an emission filter 6e that transmits predetermined fluorescence.

  Excitation light from the excitation illumination optical system 12 is reflected in the direction of the objective lens 5 by the dichroic mirror 6d in the fluorescent cube 6a, and is condensed on the sample 2 by the objective lens 5. The fluorescence expressed in the specimen 2 is collected by the objective lens 2 and enters the fluorescent cube 6a, passes through the dichroic mirror 6d and the emission filter 6e in the fluorescent cube 6a, and forms an image on the imaging device 8 by the imaging lens 7. Is done. The formed fluorescent image of the specimen 2 is picked up by the image pickup device 8, processed by the image processing circuit 10 of the control device 9, stored as a fluorescent image of the sample 2 in a memory (not shown), and connected to the control device 9. It is displayed on the display device 11.

  Using the acquired phase difference image (hereinafter referred to as a specimen image) and a fluorescence image, the control device 9 extracts a portion expressing normal fluorescence by an image analysis method based on a procedure described later, and is extracted. The fluorescence expression level is measured from the fluorescence image of the region. In this way, the fluorescence measuring device 1 is configured.

  Moreover, the acquisition of the specimen image by transmitted illumination and the fluorescence image by excitation light illumination is continuously performed at time intervals that allow cell movement of the specimen 2 to be ignored.

  Note that either the sample image or the fluorescence image may be acquired first, but in the following description, the fluorescence image is acquired first, then the sample image is acquired and image analysis is performed, and the fluorescence expression level is measured. The case where it does is demonstrated.

  Hereinafter, an image analysis method for extracting a cell portion expressing normal fluorescence in the specimen 2 will be described with reference to FIGS. The image analysis is executed by the control device 9 and the image processing circuit 10.

  FIG. 2 shows an analysis flow of the image analysis method according to the embodiment. FIG. 3 is a graph showing the temporal transition of the amount of fluorescence expression from the cells in the analysis flow of the image analysis method according to the embodiment. FIG. 4 is a graph of the fluorescence from the cells in the analysis flow of the image analysis method according to the embodiment. FIG. 5 is a graph showing the fluorescence intensity and the expression frequency. FIG. 5 shows a case where the fluorescence from the cell is masked with the fluorescence luminance threshold Lc in the analysis flow of the image analysis method according to the embodiment. FIG. 7 shows the case where the texture feature amount in the cell is used in the analysis flow of the image analysis method according to the embodiment, when the cell area is masked by the cell area in the analysis flow of the image analysis method according to the embodiment. FIG. 8 is a diagram schematically showing each cell whose fluorescence expression level extracted in the procedure in the analysis flow of the image analysis method according to the embodiment is to be measured.

  Usually, cells that show normal fluorescent expression have the feature that the amount of fluorescent protein introduced is not excessive and the expression remains at a specific site. Therefore, the accuracy is improved by extracting the fluorescent protein that is normally expressed using at least two pieces of information, that is, the fluorescence luminance value and the spatial arrangement. If the amount of fluorescent protein introduced becomes excessive, the luminance value of the fluorescent image will increase, so taking advantage of this property to exclude cells whose luminance value exceeds the preset luminance upper limit from the extraction target, the normal amount of fluorescence expression can be reduced. It becomes possible to measure.

  Moreover, since the normal expression of fluorescence remains at a specific site, the area occupied in the cell is determined. When cells are observed at a certain magnification, there is a typical area of the nucleus and a typical area of the cytoplasm, so it is normal to exclude areas that deviate significantly from the area where the target protein should be expressed The reliability expressed in the site can be given. Since only the site where the target protein is expressed is extracted, it is not necessary to obtain detailed spatial information such as the shape and structure of the cells, and measurement is possible when a large number of cells are included in one field of view. In addition, since a large number of cells can be put in one observation field of view, it is possible to repeatedly perform image analysis from one screen and re-evaluate the measured fluorescence expression level.

  Hereinafter, the analysis flow will be described step by step with reference to FIG.

(Step S1) Fluorescence image acquisition According to the operation description of the fluorescence detection apparatus 1 described above, a fluorescence image is acquired and stored in a memory.

(Step S2) Response Correction Since the luminance information obtained from the fluorescence image includes illumination unevenness and detection sensitivity unevenness in addition to the fluorescence amount of the cells, response correction is performed on the acquired fluorescence image, and the response correction is performed. The fluorescence image of is stored in memory. Here, the response correction is to perform a process of correcting the unevenness derived from the apparatus from the fluorescence image in order to accurately calculate the amount of fluorescence from the cells. For example, a fluorescent screen that shines uniformly when illuminated with uniform illumination light is observed with this device in advance, and calibration values such as illumination unevenness and detection sensitivity unevenness derived from the device are acquired, and the calibration value is converted to a fluorescence image luminance value. To remove unevenness derived from the apparatus.

(Step S3) Obtaining Sample Image (Phase Difference Image) Similarly, the fluorescence detection device 1 obtains the sample image and stores it in the memory according to the above description.

(Step S4) Luminance Correction Luminance correction is performed to make the sample image a predetermined image, and the sample image after the luminance correction is stored in the memory. Here, the luminance correction is different from the response correction described above, and refers to performing a process of aligning the apparent luminance over the entire screen. For example, if luminance unevenness does not occur locally, the overall luminance unevenness can be corrected simply by taking a difference from the blurred image, so that each image can be corrected in real time. However, the response correction can correspond to the absolute value, whereas the luminance correction calculates the absolute value from the average luminance of a wide area (deviated size) of each image.

(Step S5) Region selection by fluorescence luminance value Lc, which is a threshold luminance value corresponding to a fluorescent protein to be introduced into a cell in advance, is obtained in the following procedure.

  In the fluorescence image, cells showing normal expression have a gentle fluorescence amount with the luminance value Lc as the upper limit, whereas cells showing excessive (abnormal) fluorescence expression increase monotonously. It continues to light up or shows a sudden brightening after passing through a stagnant region near Lc. Therefore, it becomes possible to estimate the overexpression timing (time t) from the time transition of the average expression level obtained by dividing the fluorescence expression level as shown in FIG. 3 by the total area of the expression region. For example, the time tc at which the transition of the total fluorescence expression level or the average expression level changes to the rapid brightening as shown in FIG. 3 is estimated from the time differential value.

  Alternatively, the overexpression amount can be obtained as a threshold luminance value Lc by, for example, discriminant analysis from the luminance frequency distribution (see FIG. 4) at time t. In FIG. 4, curve a represents the frequency distribution of the luminance values of the normal fluorescence expression site, curve b represents the frequency distribution of the luminance values of the abnormal fluorescence expression site, and curve c represents the total luminance value combining both a and b. Each frequency distribution is shown.

  Alternatively, as a simpler method, a difference in luminance frequency distribution is obtained from two fluorescent images before and after the time t obtained from FIG. 3 (t−Δt and t + Δt), and the normal distribution is applied only by increasing values. Therefore, it is possible to obtain the threshold luminance value Lc by taking an average value −σ (σ: variance value).

  Using the threshold luminance value Lc obtained by the method as described above, as shown in FIG. 5, processing is performed on cells (blacked portions in the figure) in which fluorescence is expressed in the fluorescent image that is equal to or higher than the threshold luminance value Lc. A mask for creating the fluorescence expression level as a target of measurement processing in other areas is created by excluding the target from the target.

  At this time, the area to be excluded from the processing target is the mask area because the leakage of light occurs in the pixels around the target pixel of the threshold luminance value Lc in the pixels of the image sensor 8 that captures the fluorescent image. It is set to be expanded to the full width of (Point Spread Function: abbreviation of point image response function). As an expansion method for this region, an expansion process for expanding the area by one adjacent pixel, or an expansion process for labeling each overexpressed cell and expanding the shape while retaining the shape can be employed.

(Step S6) Area selection by area A luminance distribution of noise is created from a fluorescence luminance value of a region in which fluorescence expression is not observed from a fluorescent image acquired in the early stage of fluorescence observation, and a noise distribution is obtained assuming a normal distribution.

  Next, in the fluorescence image area of the mask, an area having a fluorescence luminance value that is twice or more the variance value of the noise is extracted. Here, only the area where the fluorescence is expressed is calculated, and when the value is out of the range of the area suitable for the expression site, it is excluded from the processing target. As shown in FIG. 6, a region whose area is out of a predetermined range is excluded from the processing target, and a mask is created with the other region as the processing target. In the figure, (> Super) indicates a region where the area exceeds the upper limit value Super, and (> Slow) indicates a region where the area exceeds the lower limit value Slow, and the mask is rewritten so that this region is excluded from the measurement processing target. Set.

  In the steps so far, the mask by the threshold luminance value Lc and the process exclusion target mask are applied, and the measurement process target area is narrowed.

(Step S7) Candidate Area Mask By executing the above steps S5 and S6, a mask for the area to be measured is completed.

(Step S8) Candidate Area Extraction A cell image corresponding to the mask is extracted from the sample image using the mask set in step S7.

(Step S9) Texture Analysis In the extracted cell image, the texture feature amount in the cell is calculated by the control device 9. For example, as shown in FIG. 7, in the case of a nucleus, the area ratio between the nucleolus and other regions and the spatial frequency of light and shade are feature quantities, and when expressed in the cytoplasm, a bright halo region generated by the cell wall The area ratio of other regions and the frequency distribution of luminance values are taken as special amounts.

(Step S10) Discrimination Based on Texture Feature Amounts in which the feature amount calculated in step S9 is out of the predetermined range are excluded from the processing target.

(Step S11) Normal Expression Area Mask Using this added feature amount as a discrimination criterion, it is estimated to which part of the cell the area extracted from the fluorescence image belongs, and only those that are expressed in the correct part are extracted. Create a mask. In this way, the areas indicated by a and b in FIG. 7 are excluded, and the normal expression area mask A in the fluorescence image shown in FIG. 8 is determined.

(Step S12) Normal Expression Fluorescence Amount Extraction A normal expression region mask is applied to the fluorescence image, and the fluorescence expression amount of the extracted fluorescence region is calculated from the luminance value of the fluorescence image.

(Modification of the embodiment)
Next, an image analysis method according to a modification of the above-described embodiment for extracting a cell part expressing normal fluorescence in the specimen 2 will be described with reference to FIGS. The image analysis is executed by the control device 9 and the image processing circuit 10 as in the above embodiment.

  FIG. 9 shows an analysis flow of a modified image analysis method. FIG. 10 shows a state in which the fluorescence from the cell in the analysis flow of the modified image analysis method is masked with the threshold luminance threshold Lc, and FIG. 11 shows the cell image in the analysis flow of the modified image analysis method. FIG. 12 is a diagram schematically showing each cell site where the fluorescence expression level extracted in the procedure of the analysis flow of the image analysis method of the modification is to be measured when the mask is applied with the cell area.

  The image analysis method according to the modification is an image analysis method in which the texture feature amount processing in the first embodiment is omitted. The same reference numerals are given to the same components as those in the above embodiment, and the description thereof is omitted. Processing different from this flow will be described.

  Hereinafter, each step will be described according to the analysis flow of FIG.

  The steps from (Step S1) to (Step S5) are the same as the analysis flow of the above embodiment, and the description is omitted. FIG. 10 is also equivalent to FIG. 5 of the above embodiment, and the description is omitted.

(Step S16) Confirmation of expression region The region selected by the mask set by excluding the region exceeding the threshold luminance value Lc from the processing target is determined. Here, the candidate region is selected only from the luminance value, which corresponds to step S7 in FIG. 2 in which the region is extracted without using the area. However, the area below the noise level used in step S6 is excluded.

  The expression region is determined on the basis of the area ratio between the fluorescence expression region and the fluorescence non-expression region in the specimen image or the luminance ratio of the fluorescence through the following steps.

(Step S17) Segmentation into Single Cells Characterized by a halo (indicated by a broken line in FIG. 11) that occurs in a region where the phase difference corresponding to the cell wall is severe in the sample image, as shown in FIG. Separate cells individually.

(Step S18) Extraction of Cell Region Containing Fluorescence Expression Then, the fluorescence expression region is matched from the specimen image and the fluorescence image of the individually cut cells, and only the cells containing the fluorescence expression are extracted. For example, the region indicated by the symbol X in FIG. 11 is excluded because the fluorescent region is not included in a single cell.

(Step S19) Selection by expression / non-expression region area ratio Since the area ratio between the expression region and the non-expression region in the cell has a specific value depending on the expression site, the normal expression region and the abnormal expression region are determined based on this. To determine. For example, the area of reference Y in FIG. 11 is excluded because the ratio (area ratio) occupied by the fluorescence expression area in a single cell is not within a predetermined value. Based on this determination, the normal expression region mask A of the region for acquiring the fluorescence expression region as shown in FIG. 12 is determined.

(Step S20) Normal Expression Fluorescence Amount Extraction A normal expression region mask is set in the fluorescence image, and the fluorescence expression amount in the set region is measured from the fluorescence image.

  As described above, since the image analysis method of the modification uses the area and the band-like luminance value as the determination criterion, the determination can be made even if the spatial resolution is not so high that the texture structure such as the nucleolus can be observed. Is also a strong treatment method.

  By executing each step of the image analysis method according to the embodiment or the image analysis method according to the modification of the embodiment, the abnormal fluorescence expression region may be removed, and only the normal fluorescence expression region may be subjected to measurement processing. And a more accurate fluorescence expression level can be detected. In addition, it becomes possible to quantitatively and accurately evaluate the degree of action of the fluorescent reagent.

  Moreover, the optimal timing of cell observation can be known by applying the above steps to time-series acquired images (fluorescence images and specimen images).

  Further, by analyzing the histogram of the luminance value of each fluorescent image in time series, the threshold luminance value Lc can be determined more accurately from the frequency distribution shown in FIG. Moreover, the temporal transition of the fluorescence expression level can be detected.

  Since the time-series fluorescence image and the sample image are stored in the memory, the optimum threshold luminance value Lc is determined, or each mask is changed as appropriate, and the process for each image is repeated. Is possible. By repeating the trial, the threshold luminance value Lc or each mask setting can be optimized, and a more accurate fluorescence expression amount can be obtained. It is also possible to set the mask formed in each step to a fluorescence image and measure the amount of fluorescence expression, thereby evaluating the accuracy of mask setting etc. from the change in the measured value for each mask. is there. Which mask is used to adopt the fluorescence expression level is appropriately examined and used by the measurer.

  As described above, according to the image analysis method according to the above-described embodiment and the fluorescence detection apparatus having the image analysis method, by extracting only fluorescent proteins that show normal expression, excessive introduction of proteins that are not originally meaningful It is possible to remove cells that have been killed by or that show expression at abnormal sites. For this reason, it is possible to obtain accurate fluorescent protein introduction efficiency that has not been obtained so far. Further, since not only the fluorescence observation image but also the bright field observation image (phase difference image) is acquired at the same time, the fluorescence expression site from the initial stage of introducing the fluorescent protein in which the expression of the fluorescence, which HCS is not good at, is hardly seen. Thus, it is possible to obtain the transition of the fluorescence expression level from the initial stage of the fluorescence expression.

  The above-described embodiment is merely an example, and is not limited to the above-described configuration and shape, and can be appropriately modified and changed within the scope of the present invention.

1 shows a schematic configuration of a fluorescence detection apparatus according to an embodiment. The analysis flow of the image analysis method concerning embodiment is shown. The time transition of the fluorescence expression level from the cell in the analysis flow of the image analysis method concerning embodiment is shown. The fluorescence intensity and expression frequency of fluorescence from cells in the analysis flow of the image analysis method according to the embodiment are shown. A time when the fluorescence from the cell in the analysis flow of the image analysis method according to the embodiment is masked with the fluorescence luminance threshold value Lc is shown. The time when a cell area is masked in the cell image in the analysis flow of the image analysis method according to the embodiment is shown. The time of estimating a cell site | part using the texture feature-value in a cell in the analysis flow of the image analysis method concerning embodiment is shown. It is a figure which shows typically the normal expression area | region mask containing the cell which should measure the fluorescence expression level extracted by the said procedure in the analysis flow of the image analysis method concerning embodiment. The analysis flow of the image analysis method of the modification of embodiment is shown. A time when the fluorescence from the cell in the analysis flow of the image analysis method of the modification is masked with the threshold luminance threshold value Lc is shown. The time when a cell area is masked in the cell image in the analysis flow of the image analysis method of the modification is shown. It is a figure which shows typically the normal expression area | region mask containing the cell which should measure the fluorescence expression level extracted in the procedure in the analysis flow of the image analysis method of a modification.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Fluorescence detection apparatus 2 Sample 3 Stage 4 Transmission illumination optical system 5 Objective lens 6 Filter turret 7 Imaging lens 8 Imaging apparatus 9 Control apparatus (PC)
DESCRIPTION OF SYMBOLS 10 Image processing circuit 11 Monitor 12 Excitation illumination optical system 13 Phase difference filter A Normal expression area mask

Claims (12)

A specimen image and a fluorescence image of the specimen are continuously acquired by a detector, and less than a threshold luminance value which is an upper limit value of a normal fluorescence expression amount in a cell unit of the specimen into which a predetermined fluorescent protein is introduced from the fluorescence image A first region of the cell unit expressing the fluorescence of
A second region in which the ratio of the fluorescence expression area to the fluorescence non-expression area is within a predetermined range in each cell unit of the first region is extracted, and the texture feature amount of each cell in the second region in the sample image Extracting the third region of the fluorescent image in which is within a predetermined range. A specimen image and a fluorescence image of the specimen are continuously acquired by a detector, and less than a threshold luminance value which is an upper limit value of a normal fluorescence expression amount in a cell unit of the specimen into which a predetermined fluorescent protein is introduced from the fluorescence image A first region of the cell unit expressing the fluorescence of
Extract only the cells that contain the first region from the cells of the specimen image, and select the second region of the fluorescence image in which the ratio of the fluorescence expression area to the fluorescence non-expression area is within a predetermined range within each cell unit. An image analysis method characterized by extracting.   The image analysis method according to claim 1, wherein the upper limit value is determined from a time-series transition of an average expression level of fluorescence in the fluorescence image.   The image analysis method according to claim 1, wherein the upper limit value is determined from a luminance frequency distribution obtained from the fluorescence image. The fluorescence detection apparatus characterized by performing the image-analysis method of any one of Claims 1-4. Detection means having a spatial resolution for continuously acquiring a specimen image and a fluorescence image of the specimen;
Threshold setting means for setting a threshold luminance value of fluorescence from the fluorescence image;
A first region extracting means for extracting a first region of a cell unit expressing fluorescence less than the threshold luminance value from the fluorescence image;
A second region extracting means for extracting a second region in which the ratio of the fluorescence expression area and the fluorescence non-expression area is within a predetermined range within each cell unit in the first region;
A third region extracting means for extracting a third region in which the texture feature amount of each cell in the second region in the sample image is within a predetermined range;
Measuring means for measuring the fluorescence expression amount from the fluorescence image corresponding to the third region,
The fluorescence detection apparatus according to claim 1, wherein the threshold luminance value is an upper limit value of a normal fluorescence expression amount in the cell unit of the sample into which a predetermined fluorescent protein is introduced. A determination means for determining whether or not the third region is a desired cell site;
The fluorescence detection apparatus according to claim 6, wherein the measurement unit measures a fluorescence expression level from the fluorescence image corresponding to the determined cell site.   The fluorescence detection apparatus according to claim 6 or 7, wherein the measurement unit measures a fluorescence expression amount from the fluorescence image corresponding to the second region. Detection means having a spatial resolution for continuously acquiring a specimen image and a fluorescence image of the specimen;
Threshold setting means for setting a threshold luminance value of fluorescence from the fluorescence image;
A first region extracting means for extracting a first region of a cell unit expressing fluorescence less than the threshold luminance value from the fluorescence image;
Only cells that contain the first region in the specimen image are extracted, and a second region in which the ratio of the fluorescence expression area to the fluorescence non-expression area is within a predetermined range in each cell unit in the first region is extracted. Second region extracting means for performing,
Measuring means for measuring the fluorescence expression amount from the fluorescence image corresponding to the second region,
The fluorescence detection apparatus according to claim 1, wherein the threshold luminance value is an upper limit value of a normal fluorescence expression amount in the cell unit of the sample into which a predetermined fluorescent protein is introduced.   The fluorescence detection apparatus according to any one of claims 6 to 9, wherein the upper limit value is determined from a time series transition of an average expression level of fluorescence in the fluorescence image.   The fluorescence detection apparatus according to any one of claims 6 to 9, wherein the upper limit value is determined from a luminance frequency distribution obtained from the fluorescence image. The fluorescence detection apparatus according to claim 6, wherein the measurement unit measures a fluorescence expression amount from the fluorescence image corresponding to the first region.

Images