WO2018139242A1

WO2018139242A1 - Image analysis system, image analysis method, and program

Info

Publication number: WO2018139242A1
Application number: PCT/JP2018/000893
Authority: WO
Inventors: 靖之祖父江; 靖裕間宮
Original assignee: パナソニック株式会社
Priority date: 2017-01-27
Filing date: 2018-01-16
Publication date: 2018-08-02

Abstract

Provided are an image analysis system, image analysis method, and program that make it possible to detect a specific substance from a liquid sample including a plurality of types of substances even if the brightness of the overall image changes. This image analysis system (100) is provided with a creation unit (111), determination unit (112), and decision unit (113). The creation unit (111) extracts a target area from an image that comprises a set of pixels that is smaller than the number of pixels in the image and includes pixels having a brightness difference with surrounding pixels of a prescribed value or greater and the surrounding pixels. The creation unit (111) creates a brightness histogram associating the brightnesses included in the target area and the frequencies of occurrence of the brightnesses. The determination unit (112) determines a brightness threshold from the brightness histogram. The decision unit (113) uses the brightness threshold to determine whether a specific substance is at least in the target area extracted by the creation unit (111).

Description

Image analysis system, image analysis method and program

The present invention relates to an image analysis system, an image analysis method, and a program, and more particularly, to an image analysis system, an image analysis method, and a program for analyzing an image and detecting a region where a specific substance exists from the image.

Conventionally, there is a system that detects fluorescence (specific substance) of a biological substance that is fluorescently stained among a plurality of types of substances from an image (for example, Patent Document 1).

The system described in Patent Literature 1 creates a graph in which brightness and the frequency of appearance of brightness are associated with each other in an image, and the biological material that is fluorescently stained using a predetermined threshold for the created graph. Detect fluorescence. This makes it possible to detect an area where a specific substance exists from the image.

* The brightness of the entire image may vary depending on conditions such as the environment during measurement. Therefore, when a predetermined threshold value is used, there is a possibility that the fluorescence of the biological material (specific substance) may not be detected depending on the brightness of the entire image.

Japanese Patent No. 5906623

Therefore, the present invention has been made in view of the above reasons, an image analysis system capable of detecting a specific substance from a liquid sample containing a plurality of kinds of substances even when the brightness of the entire image has changed, An object is to provide an image analysis method and program.

The image analysis system according to an aspect of the present invention includes a creation unit, a determination unit, and a determination unit. The creation unit includes a pixel having a large difference from the brightness of peripheral pixels from the image and the peripheral pixel, and extracts a region composed of a collection of pixels smaller than the number of pixels of the image as a target region, A luminance histogram is created in which the luminance included in the target region is associated with the frequency at which the luminance appears. The determination unit determines a luminance threshold value from the luminance histogram. The determination unit determines whether or not a specific substance exists in at least the target region extracted by the creation unit, using the luminance threshold value.

The image analysis method according to an aspect of the present invention includes a creation step, a determination step, and a determination step. The creation step includes a pixel having a large difference from the brightness of peripheral pixels from the image and the peripheral pixel, and extracts a region composed of a collection of pixels smaller than the number of pixels of the image as a target region, A luminance histogram is created in which the luminance included in the target region is associated with the frequency at which the luminance appears. The determining step determines a brightness threshold value from the brightness histogram. The determination step determines whether or not a specific substance is present in at least the target region extracted in the creation step using the luminance threshold value.

The program according to one aspect of the present invention is a program for causing a computer to function as the image analysis system.

FIG. 1 is a block diagram showing a configuration of an image analysis apparatus according to Embodiment 1 of the present invention. FIG. 2 is a configuration diagram of a detection apparatus that includes the image analysis apparatus as described above and inspects a liquid sample using a disk. FIG. 3 is a plan view of the disk according to Embodiment 1 of the present invention. FIG. 4A is a perspective view of the above disk. FIG. 4B is a partially broken perspective view of the above disk. FIG. 4C is an enlarged view of a main part C of FIG. 4B. FIG. 4D is a perspective view of the same disk as seen from below the main part. FIG. 5 is an exploded perspective view of the disk. FIG. 6A is a perspective view of the same disk as seen from the upper side of the disk body. FIG. 6B is a perspective view of the same disk as seen from the lower side of the disk main body. 6C is a cross-sectional view taken along the line X1-X1 of FIG. 6A. FIG. 6D is an enlarged view of a main part of FIG. 6C. 7A to 7D are diagrams for explaining the flattening process performed by the image analysis apparatus same as above. FIG. 8 is a diagram showing an example of a luminance histogram obtained after flattening processing performed by the image analysis apparatus same as above. FIG. 9 is a diagram for explaining extraction of a target area performed by the image analysis apparatus same as above.

Each embodiment and modification described below are only examples of the present invention, and the present invention is not limited to each embodiment and modification. Even if other than these embodiments and modifications, various modifications can be made according to the design and the like as long as they do not depart from the technical idea of the present invention.

(Embodiment 1)
An image analysis system, an image analysis method, and a program according to this embodiment will be described with reference to FIGS.

The image analysis system 100 according to the present embodiment is provided in a detection system for detecting a specific substance from a liquid sample containing a plurality of types of substances.

As shown in FIG. 2, the detection system as a computer system includes a detection device 70 that detects a specific substance from a liquid sample containing a plurality of types of substances stored in a disk 1 that is a liquid sample inspection disk. The An image analysis system 100 as a computer system includes an image analysis apparatus 101 that detects the presence or absence of a specific substance from an image obtained by analyzing a liquid sample.

First, the disk 1 will be described.

The disk 1 includes a disk-shaped laminated disk main body 2 and a plurality of filter cartridges 3. As shown in FIGS. 4A to 4C and 5, the laminated disk main body 2 includes a disk-shaped disk main body 4 and a disk-shaped plate 5 that is more flexible than the disk main body 4.

In the laminated disk body 2, the disk body 4 and the plate 5 are joined so as to overlap each other. For example, in the laminated disk main body 2, a disk-shaped disk main body 4 and a disk-shaped plate 5 are laminated via a joint 6. In the laminated disk main body 2, the central axis 45 (see FIG. 5) of the disk main body 4 and the central axis 56 (see FIG. 5) of the plate 5 are aligned on a straight line.

The disk main body 4 has a chamber 400 for storing a liquid sample. The disc body 4 has a first surface 41 and a second surface 42 that are opposite to each other in the thickness direction. The plate 5 is joined to the disc body 4 so as to cover the chamber 400 on the first surface 41 side of the disc body 4. The chamber 400 includes a first chamber 401 and a second chamber 402 as shown in FIGS. 3, 6B and 6C. The first chamber 401 penetrates in the thickness direction of the disc body 4, and the opening on the plate 5 side is closed by the plate 5. As a result, in the laminated disk main body 2, the first chamber 401 is open on the side opposite to the plate 5 side in the thickness direction of the disk main body 4. The second chamber 402 is formed on the first surface 41 of the disc body 4, and the side opposite to the plate 5 side in the thickness direction of the disc body 4 is closed. In the second chamber 402, the opening on the plate 5 side is closed by the plate 5. The second chamber 402 is in communication with (connected to) the first chamber 401. Here, the disc body 4 has a channel 403 (see FIGS. 4B, 4D, 6B, and 6C) that communicates with the first chamber 401 and the second chamber 402 between the first chamber 401 and the second chamber 402, respectively. Is preferred. The channel 403 is formed on the first surface 41 of the disk main body 4, and the side opposite to the plate 5 side in the thickness direction of the disk main body 4 is closed.

In the laminated disk main body 2, the space surrounded by the inner wall surface of the first chamber 401 and the plate 5 in the disk main body 4 constitutes a first well 21 (see FIGS. 3 and 4B) for storing a liquid sample. In the laminated disk main body 2, the space surrounded by the inner wall surface of the second chamber 402 and the plate 5 in the disk main body 4 is a second well 22 (see FIG. 3 and FIG. 3) that stores the liquid sample moved from the first well 21. 4B). In the laminated disk main body 2, a space surrounded by the inner wall surface of the channel 403 and the plate 5 in the disk main body 4 is a flow path 23 (see FIG. 5) through which a liquid sample passes between the first chamber 401 and the second chamber 402. 3 and 4B).

The liquid sample contains multiple types of substances. The filter cartridge 3 includes a filter 35 (see FIGS. 3, 4B, and 4D) that removes a specific substance from the liquid sample moving from the first chamber 401 to the second chamber 402. Here, “removing a specific substance” means capturing a specific substance. In short, the filter 35 includes a porous structure that captures a specific first substance from the liquid sample and passes the specific second substance. The filter cartridge 3 is placed in the first chamber 401 of the disc body 4. Here, in the disk 1, the filter cartridge 3 is fitted into the first well 21 of the laminated disk main body 2.

The filter cartridge 3 includes a case 30 that holds the filter 35. The case 30 has substantially the same shape as the first well 21 when viewed from the thickness direction of the laminated disk main body 2. The case 30 has a shape in which the width gradually increases with increasing distance from the center of the laminated disk body 2 in the radial direction of the laminated disk body 2 when viewed from the thickness direction of the laminated disk body 2. The case 30 has an opening 320 (see FIG. 4D) on one surface of the second chamber 402 side. In the filter cartridge 3, the filter 35 is disposed so as to close the opening 320 of the case 30. The filter 35 is fixed to the case 30 with, for example, an adhesive. In the filter cartridge 3, the space surrounded by the case 30 and the filter 35 constitutes a storage space 31 for liquid sample (see FIGS. 3 and 4B).

The disk 1 is used, for example, to examine the infection rate of pathogenic microorganisms (for example, malaria protozoa) to a specimen (for example, red blood cells) in a liquid biological sample (for example, human blood). The malaria parasite, for example, invades a human body when an mosquito sucks human blood, invades red blood cells in the blood, and parasitizes in red blood cells. The “infection rate” here is {[number of samples infected with pathogenic microorganisms] / [total number of samples]} × 100 [%]. The liquid sample includes at least a liquid biological sample. In the liquid sample, for example, when the biological sample is blood, the blood is preferably diluted with a diluent in order to reduce the viscosity. As the diluted solution, a solution that does not denature blood cells (erythrocytes, leukocytes) contained in the biological sample is used. As the diluent, for example, a buffer solution, an isotonic solution, a culture solution, a surfactant and the like can be used.

In the disk 1, it is preferable that a fluorescent reagent (fluorescent dye) for staining the nucleic acid of pathogenic microorganisms is disposed in the second well 22 of the laminated disk body 2. The fluorescent reagent is preferably arranged by, for example, a freeze-drying method or a spin coating method. Thereby, the disc 1 can fluorescently label the nucleic acid of the pathogenic microorganism that is parasitic on the specimen (red blood cells) in the liquid sample moved to the second well 22. The nucleic acid stained with the fluorescent reagent emits fluorescence when excitation light is irradiated from the outside. The fluorescent reagent for staining the nucleic acid of the pathogenic microorganism may be a powder.

In the disk 1, the filter 35 is configured to pass red blood cells that are specific second substances (specimens) and to capture white blood cells that are specific first substances. In other words, the filter 35 is configured to function as a separation unit that separates red blood cells and white blood cells and extracts red blood cells. Therefore, the disk 1 can extract red blood cells from a biological sample.

Fluorescent reagents for staining nucleic acids of pathogenic microorganisms are materials that can also stain leukocytes. However, in the disk 1, leukocytes in the liquid sample placed in the first chamber 401 are captured by the filter 35. Therefore, in the disc 1, it is possible to prevent the white blood cells contained in the liquid sample put in the first chamber 401 from being stained with the fluorescent reagent.

The filter 35 in the filter cartridge 3 exists between the storage space 31 and the second chamber 402 in the radial direction of the disc body 4. In the disk 1, the filter cartridge 3 is placed in the first chamber 401 of the disk body 4, so that the liquid sample in the storage space 31 of the filter cartridge 3 can be regarded as the liquid sample placed in the first chamber 401. In the disk 1, since the filter 35 in the filter cartridge 3 is between the storage space 31 and the second chamber 402, red blood cells in the liquid sample placed in the storage space 31 are moved to the second well 22 through the filter 35. It becomes possible. The operation of putting the liquid sample into the storage space 31 is preferably performed in a state in which the filter cartridge 3 is fitted in the first well 21 of the laminated disk main body 2.

The shape of the storage space 31 of the filter cartridge 3 is U-shaped as shown in FIG. 3 when viewed from the thickness direction of the disk 1. In the filter cartridge 3, it is desirable that the injection hole 33 is provided so as to communicate with the first end of the U-shaped storage space 31 in the case 30 and the vent hole 38 is provided so as to communicate with the second end. . Thereby, when the liquid sample is injected into the storage space 31 of the filter cartridge 3, the air existing in the storage space 31 can be released from the portion other than the filter 35, so that the liquid sample can be smoothly supplied. Can be injected. The shape of the vent 38 is, for example, a circle. The vent hole 38 is preferably smaller from the viewpoint of preventing leakage of the liquid sample, and is preferably smaller than the injection hole 33.

In the disk 1, the storage space 31, the filter 35, and the second chamber 402 are arranged in this order from the center side of the disk body 4 to the outer peripheral side in a state where the filter cartridge 3 is placed in the first chamber 401. Is preferred. In short, in the disk 1, it is preferable that the storage space 31, the filter 35, and the second well 22 are arranged in this order from the center side of the laminated disk main body 2 outward in the radial direction of the laminated disk main body 2. Thereby, the liquid sample in the storage space 31 can be moved to the second well 22 through the filter 35 by the centrifugal force acting on the liquid sample when the disk 1 is rotated. In the second well 22, surface tension or the like acts on the liquid sample in addition to centrifugal force. The rotation direction of the disk 1 is clockwise (clockwise) when viewed from the upper side of the disk 1 (the second surface 42 side of the disk body 4 in the disk 1).

The shape of the laminated disk main body 2 is preferably a disk shape as in the case of optical disks (CD, DVD, etc.). A circular hole 28 is preferably formed in the center of the laminated disk body 2. The diameter of the disk 1 is 120 mm, for example.

The laminated disk main body 2 includes the disk-shaped disk main body 4 and the disk-shaped plate 5 joined to the disk main body 4 on the first surface 41 side of the disk main body 4 as described above. Here, a circular hole 48 constituting a part of the hole 28 of the laminated disk main body 2 is formed in the center of the disk main body 4. A circular hole 58 constituting a part of the hole 28 of the laminated disk main body 2 is formed at the center of the plate 5. The plate 5 includes a disk-shaped plate body 50 (see FIG. 4C). The material of the plate body 50 is, for example, a transparent resin. The plate body 50 has a front surface 51 and a back surface 52 that are opposite to each other in the thickness direction. The front surface 51 of the plate main body 50 is preferably formed with a spiral track for following the beam-like light incident through the back surface 52 of the plate main body 50 in the same manner as the optical disc. A track is a groove. The track is formed in a spiral shape from the center to the outer periphery of the plate body 50. Address information is continuously recorded on the track. Thereby, in the plate body 50, the position can be specified by the address information. Therefore, for example, the position information of the second well 22 in the plane of the disk 1 is specified by the address information. As in the case of a CD or DVD, the track 1 is scanned with light to reproduce address information. The light is excitation light. The wavelength of the excitation light is preferably 400 nm to 410 nm, for example, and more preferably 405 nm. The track depth is, for example, 50 nm.

The plate 5 further includes a dielectric film 54 (see FIG. 4C) formed on the surface 51 of the plate body 50. The dielectric film 54 is, for example, a ZnS—SiO ₂ film. The dielectric film 54 is formed so as to cover the track. The dielectric film 54 is configured to reflect a part of the excitation light for tracking and transmit most of the remaining part. The reflectance of the dielectric film 54 with respect to the excitation light is, for example, 5% or more and 20% or less. For example, the reflectance of the dielectric film 54 with respect to the fluorescence is preferably less than or equal to the reflectance of the dielectric film 54 with respect to the excitation light. In the disk 1, a reflection surface 55 (see FIG. 4C) that reflects the excitation light incident on the back surface 52 of the plate body 50 is configured by an interface between the dielectric film 54 and the plate body 50.

The specimen in the liquid sample sent from the first well 21 to the second well 22 through the filter 35 in the disk 1 is inspected by, for example, a detection device 70 as shown in FIG.

The detecting device 70 includes, for example, an optical system similar to an optical pickup device for an optical disc, and the operation thereof is also the same. The optical system of the detection device 70 includes a semiconductor laser 71, a polarization beam splitter 72, an objective lens 73, a dichroic prism 74, a fluorescence detector 75, an anamorphic lens 76, and a reflected excitation light detector 77. ing.

In addition to the optical system described above, the detection device 70 includes a holder 81, an actuator 82, a rotation device 83, a first signal calculation circuit 84, a servo circuit 85, a second signal calculation circuit 86, and image analysis. A device 101 (image analysis system 100) and an image display device 88 are provided. The rotating device 83 is a motor. The rotating device 83 is controlled by a servo circuit 85.

In the detecting device 70, after the disk 1 is set on the rotating table by the rotating device 83, a predetermined operation is started.

The optical system, the holder 81 and the actuator 82 are installed in a housing in the same manner as an existing optical pickup device used for recording / reproducing of a CD or DVD. The housing is movable in the radial direction of the disk 1 by a predetermined guide mechanism. The servo circuit 85 also controls the movement of the housing. Since this control is the same access control as that in the existing CD player or DVD player, detailed description thereof is omitted.

The semiconductor laser 71 emits light (excitation light) having a wavelength of about 405 nm. In FIG. 2, the traveling path of light is indicated by a one-dot chain line. The excitation light emitted from the semiconductor laser 71 is reflected by the polarization beam splitter 72 and enters the objective lens 73.

The objective lens 73 has a predetermined numerical aperture (Numerical Aperture) and is configured to properly converge the excitation light on the disk 1. Specifically, the objective lens 73 is configured such that excitation light incident from the polarization beam splitter 72 side converges.

The objective lens 73 is driven by the actuator 82 in the focus direction (the thickness direction of the disk 1) and the tracking direction (the radial direction of the disk 1) while being held by the holder 81. That is, the objective lens 73 is driven so as to follow the track in a state where the excitation light is focused on the reflection surface 55 (see FIG. 4C) of the disk 1. A part of the excitation light focused on the reflection surface 55 is reflected by the reflection surface 55 and most of the excitation light is transmitted through the reflection surface 55.

Fluorescence is generated when the excitation light focused by the objective lens 73 is irradiated onto a nucleic acid that is fluorescently labeled in red blood cells. The wavelength of fluorescence is different from the wavelength of excitation light. The fluorescence wavelength is preferably, for example, 440 nm to 490 nm, and more preferably 455 nm. For example, SYTO (registered trademark) Blue can be used as the fluorescent dye. Red blood cells that are not infected with malaria parasites are not fluorescently labeled, and therefore do not generate fluorescence even when irradiated with excitation light. Therefore, the detection apparatus 70 can distinguish between red blood cells infected with malaria parasites and red blood cells not infected by the presence or absence of fluorescence.

The dichroic prism 74 is configured to reflect light having a wavelength of about 405 nm and transmit light having a wavelength of about 440 to 600 nm.

Excitation light reflected by the reflecting surface 55 (hereinafter referred to as “reflected excitation light”) passes through the polarization beam splitter 72, is reflected by the dichroic prism 74, and enters the anamorphic lens 76.

The anamorphic lens 76 introduces astigmatism into the reflected excitation light incident from the polarization beam splitter 72 side. The reflected excitation light transmitted through the anamorphic lens 76 enters the reflected excitation light detector 77. The reflected excitation light detector 77 has a four-divided sensor for receiving reflected excitation light on the light receiving surface. The detection signal of the reflected excitation light detector 77 is input to the second signal calculation circuit 86.

The second signal calculation circuit 86 generates a focus error signal and a tracking error signal from the detection signal of the reflected excitation light detector 77, and also generates a wobble signal (Wobble Signal). The focus error signal is a signal indicating a deviation (focus error) between the focal position of the objective lens 73 and the disk 1. The tracking error signal is a signal indicating a deviation (tracking error) between the spot of the excitation light and the track. The wobble signal is a waveform signal corresponding to the meandering shape of the groove defined by the track. The focus error signal and the tracking error signal are generated according to the astigmatism method and the one-beam push-pull method. The wobble signal is generated based on the tracking error signal. Specifically, a wobble signal is generated by extracting a frequency component corresponding to the wobble signal from the tracking error signal. The servo circuit 85 controls the actuator 82 using the focus error signal and tracking error signal output from the second signal calculation circuit 86. The servo circuit 85 controls the rotating device 83 so that the disk 1 is rotated at a predetermined linear velocity using the wobble signal output from the second signal calculation circuit 86.

Also, the second signal calculation circuit 86 outputs reproduction data (address information) generated by demodulating the wobble signal to the image analysis apparatus 101.

Fluorescence incident on the dichroic prism 74 from the objective lens 73 side passes through the dichroic prism 74 and enters the fluorescence detector 75. The fluorescence detector 75 has a sensor that converts the received fluorescence into a detection signal composed of an electrical signal and outputs the detection signal. The detection signal of the fluorescence detector 75 is input to the first signal calculation circuit 84.

The first signal calculation circuit 84 outputs fluorescence luminance information generated by amplifying the detection signal from the fluorescence detector 75 to the image analysis apparatus 101.

The image analysis apparatus 101 is configured to display an image of the liquid sample in the second chamber 402 based on the fluorescence luminance information output from the first signal calculation circuit 84 and the address information output from the second signal calculation circuit 86. Is generated and displayed on the image display device 88. The image analysis apparatus 101 performs a predetermined process on the generated image, determines whether or not red blood cells infected with malaria parasites exist in the liquid sample, and causes the image display apparatus 88 to display the result. The image analysis apparatus 101 can be realized, for example, by causing a personal computer to execute an appropriate program. Further, the image display device 88 can be constituted by a display of a personal computer, for example.

The procedure of an example of examining red blood cells using the disk 1 and the detection device 70 will be briefly described.

After preparing blood (biological sample) collected from a patient, a liquid sample is prepared by mixing the blood and a diluent.

Thereafter, a liquid sample is put into the storage space 31 of the filter cartridge 3. When putting a biological sample in the storage space 31, for example, a pipette (Pipette), a syringe (Syringe), a capillary (Capillary) or the like is used. Here, it is preferable to put the liquid sample into the storage space 31 in a state where the filter cartridge 3 is fitted in the first well 21 of the laminated disk main body 2.

Thereafter, in the detection device 70, the disk 1 is rotated at a predetermined linear velocity for a predetermined rotation time. The detection device 70 rotates the disk 1 around the central axis 25 of the laminated disk main body 2. At this time, white blood cells in the liquid sample are captured by the filter 35 of the filter cartridge 3 and do not reach the flow path 23 and the second well 22. Therefore, in the disk 1, the red blood cells contained in the liquid sample can be moved from the first well 21 to the second well 22, and the white blood cells contained in the liquid sample can be captured by the filter 35. .

Here, the configuration of the image analysis apparatus 101 (image analysis system 100) will be described.

The image analysis apparatus 101 includes a first input unit 102, a second input unit 103, a control unit 104, and an output unit 105 (see FIG. 1). The image analysis apparatus 101 includes a CPU (Central Processing Unit) and a memory. The function of the control unit 104 is realized by the CPU executing a program stored in the memory. Here, the program is provided through an electric communication line such as the Internet or recorded in a recording medium such as a memory card, but may be recorded in advance in a memory of a computer.

The first input unit 102 receives the fluorescence luminance information output from the first signal calculation circuit 84. The second input unit 103 receives the address information output from the second signal calculation circuit 86.

The control unit 104 includes a first processing unit 110, a creation unit 111, a determination unit 112, a determination unit 113, and a second processing unit 114 (see FIG. 1).

The first processing unit 110 generates an image of the sample separated from the liquid sample. Specifically, the first processing unit 110 uses the fluorescence luminance information received from the first signal calculation circuit 84 and the address information received from the second signal calculation circuit 86 to store the information in the second well 22. Generate an image of the sample. The first processing unit 110 outputs the generated image to the image display device 88 via the output unit 105.

The creation unit 111 uses the image generated by the first processing unit 110 to create a luminance histogram G10 representing the correspondence between the luminance and the frequency at which the luminance appears in the image (see FIG. 8). Specifically, the creation unit 111 creates a saliency map from the image generated by the first processing unit 110, and using the created saliency map, a difference in luminance with surrounding pixels is a predetermined value or more. All regions (target regions) including a certain pixel are extracted. The creation unit 111 creates a brightness histogram G10 using all the extracted regions. Hereinafter, in the description, luminance means a luminance value. For example, the luminance is any value from 0 to 255.

Hereinafter, extraction of the target area will be described with reference to FIG.

The creation unit 111 acquires one or more first pixel groups B10 that are a collection of pixels with high luminance (bright pixels) from an image B1 (first image described later) generated by the first processing unit 110. For example, in the illustration of FIG. 9, the creating unit 111 acquires four first pixel groups B10. The creation unit 111 acquires a second pixel group B11 that is a collection of pixels that exist around the first pixel group B10 and have a luminance lower than that of the first pixel group B10. Here, for example, the creation unit 111 acquires the second pixel group B11 so that the outer shape of the second pixel group B11 is a square shape. The creation unit 111 extracts from the image B1 a region B20 that is a combination of the first pixel group B10 and the second pixel group B11 existing around the first pixel group B10 as a target region. That is, the creation unit 111 includes a pixel having a difference from the brightness of surrounding pixels from the image B1 to a predetermined value or more and a pixel existing around the pixel, and having a smaller number of pixels than the image B1. A body region B20 is extracted as a target region. The creation unit 111 creates a brightness histogram G10 in which the brightness of each pixel included in the area B20 (target area) is associated with the frequency at which the brightness appears.

The creating unit 111 does not create the luminance histogram G10 from the entire image B1, but creates a region (target region) having a smaller number of pixels than the number of pixels of the image B1, using the saliency map. In other words, the region B20 used for creating the histogram G10 is narrowed down to a part of the image B1. Therefore, the ratio of the number of pixels of the first pixel group to the number of pixels of the target region is larger than the ratio of the number of pixels of the second pixel group to the number of pixels of the image B1. That is, the creation unit 111 sets the target area to the image B1 so that the ratio of the number of pixels of the first pixel group to the number of pixels of the target area is larger than the ratio of the number of pixels of the second pixel group to the number of pixels of the image. Extract from Here, the first pixel group is a group of pixels in the target area whose difference from the luminance of surrounding pixels is a predetermined value or more. The second pixel group is a group of pixels in the entire image having a difference from the luminance of surrounding pixels equal to or greater than a predetermined value.

Note that the creation unit 111 may use a region that is a part of the region 20 and that includes a part of the first pixel group B10 and a part of the second pixel group B11 as a target region. For example, the creation unit 111 may create a luminance histogram using the region B30, which is a half of the region 20, as a target region.

Further, in the present embodiment, the creation of the luminance histogram performed by the creation unit 111 may be applied to an image whose brightness is reversed. In this case, the creation unit 111 acquires a collection of pixels with low luminance (dark pixels) as the first pixel group. The creation unit 111 acquires a group of pixels that exist around the first pixel group and have a luminance higher than that of the first pixel group as the second pixel group. The creation unit 111 extracts, from the image B1, an area obtained by combining the first pixel group and the second pixel group present around the first pixel group as a target area. The creation unit 111 sets a target region as a region including the first pixel group and the second pixel group existing around the first pixel group, and the luminance of each pixel included in the target region and the luminance appear. A luminance histogram is created in association with the frequency of performing.

The determining unit 112 uses the luminance histogram G10 generated by the generating unit 111 to determine a luminance threshold value used for determining the presence or absence of a specific substance (malaria parasite) in the image.

The determination unit 113 determines whether or not a specific substance (malaria parasite) is present in the image created by the first processing unit 110 using the brightness threshold value determined by the determination unit 112. Specifically, the determination unit 113 determines whether or not there is an area that includes a luminance equal to or higher than a luminance threshold in the image. When determining that it exists, the determination unit 113 determines that a specific substance (malaria parasite) is present in the image.

The second processing unit 114 obtains the number of specific substances present in the image when a specific substance (malaria parasite) is present in the image based on the determination result of the determination unit 113. For example, the second processing unit 114 obtains the number of regions in the image that include a luminance equal to or higher than the luminance threshold. The second processing unit 114 outputs the result of the presence / absence of the specific substance and the number of the substance when the specific substance exists to the image display device 88 via the output unit 105.

The output unit 105 outputs the image generated by the control unit 104 and the result of the presence or absence of red blood cells infected with the malaria parasite to the image display device 88. Thereby, the image display apparatus 88 can display the result about the image produced | generated by the image analysis apparatus 101, the presence or absence of the red blood cell infected with the malaria parasite.

Next, the operation of the image analysis apparatus 101 will be described.

The first processing unit 110 generates an image of the liquid sample in the second well 22 using the fluorescence luminance information and the address information.

The creation unit 111 creates a saliency map using the image generated by the first processing unit 110. For each pixel included in the image generated by the first processing unit 110, the creation unit 111 obtains an index for determining whether or not the difference in luminance from surrounding pixels is equal to or greater than a predetermined value. The creation unit 111 compares the index obtained for each pixel with a predetermined threshold value, and creates a saliency map from the result. Specifically, the creation unit 111 smoothes the image (first image) generated by the first processing unit 110 and generates a second image. The creation unit 111 subtracts the second image from the first image to generate a difference image that represents a difference in gray value (luminance) of each pixel. Here, each pixel of the difference image serves as an index. The creation unit 111 extracts, from the first image, an area (target area) corresponding to an area where the difference in luminance from surrounding pixels is equal to or greater than a predetermined value from the difference image. This creates a saliency map.

As described above, the saliency map uses the difference image formed by the difference between the image generated by the first processing unit and the image obtained by blurring the image (smoothed image), and the luminance of the surrounding pixels. This is obtained by extracting a region including pixels whose difference is equal to or greater than a predetermined value.

For example, the creation unit 111 applies a Gaussian filter to each pixel of the first image generated by the first processing unit 110, and then divides the image into a plurality of sections (for example, a region of 9 × 9 pixels). To do. The creation unit 111 creates a smoothed second image for each of the plurality of sections by replacing the brightness of the entire section with the average brightness in the section.

It will be briefly described that all the regions in which the difference in luminance from surrounding pixels is greater than or equal to a predetermined value can be extracted from the first image by the above-described operation. FIG. 7A shows an example of a graph G1 representing the correspondence between the pixels arranged in a predetermined direction (for example, the horizontal direction of the image) in the first image and the luminance of the pixel. Since the fluorescently labeled portion is bright, the change in the luminance difference between adjacent pixels is large. Therefore, when there is a fluorescently labeled part in the first image, a part R1 exists in the graph G1. A graph G2 in FIG. 7B represents correspondence between pixels arranged in a predetermined direction (for example, the horizontal direction of the image) in the second image and the luminance of the pixel. In FIG. 7B, since the first image is smoothed, the difference in luminance from adjacent pixels is smaller than in the graph G1 in FIG. 7A. Therefore, in the graph G2, the difference from the region R1 in the graph G1 is smaller. A region R2 having a small change amount exists.

The graph G3 in FIG. 7C shows the correspondence between the pixels arranged in a predetermined direction (for example, the horizontal direction of the image) in the difference image, which is the difference in gray value between the first image and the second image, and the luminance of the pixel. Represents. The luminance of the pixel in the difference image is a difference value between the graph G1 and the graph G2. In a portion where the luminance difference with the surrounding pixels is significant in the first image (region R1), the luminance difference with the surrounding pixels is noticeable also in the difference image. Exists. The part R3 represents the difference between the part R1 in the graph G1 and the part R2 in the graph G2, and the part R3 has a maximum value L3. The maximum value L3 is a difference value between the maximum value L1 in the part R1 and the maximum value L2 in the part R2. The slope of the graph is almost the same in a portion other than the portion R1 in the graph G1 (for example, the portion R11 shown in FIG. 7A) and a portion other than the portion R2 in the graph G2 (eg, the portion R21 shown in FIG. 7A). The difference is almost constant. In the present embodiment, the difference value (substantially constant value) obtained in the region R11 and the region R21 is referred to as a reference value L4. The reference value L4 is a difference value obtained between the part R11 and the part R21, but may be a difference value between a part other than the part R1 and a part other than the part R2. Alternatively, the reference value L4 may be an average value of difference values between a plurality of locations other than the region R1 and a plurality of locations other than the region R2.

The creation unit 111 extracts, from the first image, a region of the first image corresponding to a portion where the difference between the graph G1 and the graph G2 is large as a target region. For example, an intermediate value between the maximum value L3 and the reference value L4 in the part R3 is set as a value L5. In this case, the creation unit 111 sets a region including a collection of pixels having a difference value equal to or greater than the value L5 from the periphery of the pixel having the maximum value L3 of the region R3 as a difference value (a hatched portion R4 illustrated in FIG. 7D) as a target region. Is extracted from the first image. Here, the value L5 is the predetermined threshold value described above.

As described above, the creation unit 111 can extract the target region by comparing the luminance (index) of each pixel of the difference image with a predetermined threshold (value L5).

As a result, the creation unit 111 has pixels that have a difference from the first pixel (image B1) generated by the first processing unit 110 that is greater than or equal to a predetermined value, and pixels that exist around the pixel. A region B20 made up of an aggregate of pixels smaller than the number of pixels of the first image can be extracted. Here, a pixel in which a difference value between the first image and the second image is a predetermined threshold value (value L5) or more is set as a pixel in which a luminance difference from surrounding pixels is a predetermined value or more. When the creation unit 111 extracts the target region, the number of pixels whose luminance difference with surrounding pixels is equal to or greater than a predetermined value and the number of pixels of other pixels are substantially equal. Extract the target area. That is, in the region B20 shown in FIG. 9, the number of pixels of the first pixel group B10 and the number of pixels of the second pixel group B11 are substantially equal. Here, in the first image, when there are a plurality of regions including pixels whose luminance difference with the surrounding pixels is a predetermined value or more, there are a plurality of graphs G3 illustrated in FIG. 7C. In this case, for example, the creation unit 111 obtains an intermediate value between the value L3 and the reference value L4 for each of the plurality of graphs G3. The creation unit 111 obtains an average value of the obtained plurality of intermediate values. The creation unit 111 extracts a target area including a portion corresponding to the calculated average value or more from the first image.

In the image (first image) generated by the first processing unit 110, when there is no pixel whose luminance difference is equal to or greater than a predetermined value, the creation unit 111 extracts the target region described above. That is, cannot create a saliency map. In this case, the image analysis device 101 causes the image display device 88 to display a message indicating that a specific substance (protozoan malaria) is not detected from the liquid sample.

The operation described below will be described on the assumption that a saliency map has been created.

The creating unit 111 creates a luminance histogram G10 representing the relationship between the luminance and the appearance frequency of the luminance, using all the extracted target regions. In the present embodiment, the appearance frequency is the number of pixels having a corresponding luminance. Here, the extracted region includes the number of pixels having a luminance difference greater than or equal to a predetermined value and the number of other pixels substantially equal. Therefore, the luminance histogram G10 created by the creation unit 111 is a multimodal histogram including the first peak G11 and the second peak G12, as shown in FIG. Note that the appearance frequency may be a ratio with respect to the total number of pixels in all target regions for pixels having the corresponding luminance.

The determining unit 112 determines the brightness threshold using the brightness histogram G10. Specifically, the determination unit 112 obtains the minimum value of the frequency between the first peak part G11 and the second peak part G12 that are represented by the luminance histogram G10 in FIG. The corresponding luminance is set as the luminance threshold. Here, how to obtain the minimum value will be briefly described. The determination unit 112 selects the brightness P1 as a temporary threshold. The determination unit 112 selects a first comparative luminance (for example, luminance P2) smaller than the luminance P1 and a second comparative luminance (for example, luminance P3) larger than the luminance P1. The determination unit 112 determines the frequency for the temporary threshold (for example, the frequency Q1 for the luminance P1), the frequency for the first comparative luminance (for example, the frequency Q2 for the luminance P2), and the frequency for the second comparative luminance (for example, the frequency Q3 for the luminance P3). ). If the comparison result shows that there is a frequency smaller than the frequency for the temporary threshold, the determination unit 112 sets the luminance of the frequency as a new temporary threshold. In the example of FIG. 8, when the temporary threshold value is set to the luminance P1, the luminance P3 having the frequency Q3, which is smaller than the frequency Q1 with respect to the luminance P1, is set as a new temporary threshold value. The determination unit 112 repeats this operation until both the frequency for the first comparison luminance and the frequency for the second comparison luminance become larger than the frequency for the temporary threshold. Thereby, the determination part 112 can obtain | require the local minimum value. The determination unit 112 sets the luminance (temporary threshold) for the obtained minimum value as the luminance threshold.

Note that when the determination unit 112 obtains a temporary threshold in which both the frequency of the first comparison luminance and the frequency of the second comparison luminance are larger than the frequency of the temporary threshold, a value smaller than the temporary threshold and a larger value are respectively determined. It is preferable to determine the brightness threshold by acquiring the number (for example, 15). In this case, the determination unit 112 determines an average value (first average value) of frequencies corresponding to 15 values smaller than the temporary threshold and an average value (second average) of frequencies corresponding to 15 values larger than the temporary threshold. Value). The determination unit 112 determines the temporary threshold as the luminance threshold when the frequency with respect to the temporary threshold is smaller than both the first average value and the second average value.

The determination unit 113 determines whether or not a specific substance is present in the first image created by the first processing unit 110 using the brightness threshold value determined by the determination unit 112. Specifically, the determination unit 113 determines whether or not there is a collection (extraction region) of pixels whose luminance is greater than or equal to the luminance threshold in the image generated by the first processing unit 110. Here, the extraction region is a collection of pixels that are continuous in the vertical and horizontal directions of the first image. Among pixels arranged in a predetermined direction in an image, there is a pixel whose luminance is smaller than the luminance threshold due to noise or the like between two collections of pixels that are equal to or higher than the luminance threshold (first collection and second collection). There are things to do. When there are a predetermined number or less of pixels whose luminance is smaller than the luminance threshold due to noise or the like between the first collection and the second collection, the determination unit 113 refers to the first collection and the second collection. It is considered as one gathering.

The second processing unit 114 obtains the number of extraction areas based on the determination result of the determination unit 113. The second processing unit 114 outputs the result of the presence / absence of the specific substance and the number of the substance when the specific substance exists to the image display device 88 via the output unit 105.

When the red blood cells are infected with malaria parasites, the image generated by the first processing unit 110 includes a region where red blood cells are present, a region where malaria parasites are present, and other regions (background regions). . Since the malaria parasite is fluorescently labeled, the image generated by the first processing unit 110 is displayed brightly. Therefore, the luminance of the red blood cells, the background region, and the malaria parasite displayed in the image increases in this order. Therefore, the malaria parasite is displayed by extracting from the image an area including a pixel whose luminance difference with a surrounding pixel is equal to or greater than a predetermined value from the saliency map by the above-described operation. Regions can be extracted.

Also, even when the brightness of the entire image generated by the first processing unit 110 changes, the order of the luminance of the red blood cells, the background region, and the malaria parasite does not change. Therefore, by creating a saliency map, the possibility of extracting a target area including a malaria parasite increases.

In addition, the target area is extracted so that the number of pixels whose luminance difference from surrounding pixels is equal to or greater than a predetermined value and the number of other pixels are substantially equal. Therefore, in the luminance histogram G10, the appearance frequency of the luminance indicating the malaria parasite (second peak portion G12) and the appearance frequency of the luminance indicating other than the malaria parasite (first peak portion G11) are equal. In other words, a region where the malaria parasite is present can be distinguished from other regions. Therefore, it is possible to easily search for the minimum value of the frequency between the first peak part G11 and the second peak part G12.

In the present embodiment, the creation unit 111 uses the saliency map to determine all regions including pixels whose luminance difference from the surrounding pixels is equal to or greater than a predetermined value from the image generated by the first processing unit 110. Although it was set as the structure extracted, it is not limited to this structure. The creation unit 111 may extract at least one region including a pixel having a luminance difference equal to or greater than a predetermined value from the image generated by the first processing unit 110 using a saliency map.

In the present embodiment, the determination unit 113 is configured to determine the presence or absence of a specific substance with respect to the entire image generated by the first processing unit 110, but is not limited to this configuration. The determination unit 113 may determine the presence or absence of a specific substance only for all the regions extracted by the creation unit 111.

In addition, the second processing unit 114 is configured to obtain the number of specific substances in the image generated by the first processing unit 110, but is not limited to this configuration. When a specific substance exists in the image generated by the first processing unit 110, the second processing unit 114 may obtain a location (position) where the substance exists, and the size (size) may be obtained. You may obtain | require, and you may obtain | require brightness (average brightness | luminance of the area | region where a specific substance exists). When a specific substance exists in the image generated by the first processing unit 110, the second processing unit 114 may obtain at least one of the number, location, size, and brightness.

In the present embodiment, the creation unit 111 is configured to apply the Gaussian filter to the image before the image generated by the first processing unit 110 is divided into a plurality of sections, but is not limited to this configuration. Before the image generated by the first processing unit 110 is divided into a plurality of sections, a bilateral filter may be applied to the image.

In the present embodiment, the creation unit 111 is configured to replace the brightness of the entire section with the average brightness in the section for each of the plurality of sections when generating a smoothed image. It is not limited to the configuration. For each of the plurality of sections, the creation unit 111 sets the luminance of the entire section in a specific order from the larger or smaller order when arranging the median value in the section and the brightness in the section in numerical order. A smoothed image may be generated by replacing the incoming luminance value (for example, the second smallest value, the second largest value) or the pixel luminance at a specific position.

Alternatively, for each of the plurality of sections, the creating unit 111 sets the brightness of the entire section as the larger or smaller when the average brightness, the median value of the adjacent sections, and the brightness within the section are arranged in numerical order. A smoothed image may be generated by substituting the luminance value (for example, the second smallest value, the second largest value) or the pixel luminance at a specific position in a specific order from one side.

In the present embodiment, the creation unit 111 generates a saliency map, and uses the generated saliency map to extract a region including a pixel whose luminance difference with a surrounding pixel is equal to or greater than a predetermined value. However, it is not limited to this configuration. The creation unit 111 may extract a region including a pixel having a luminance difference equal to or greater than a predetermined value by edge extraction using fast Fourier or differential edge detection. For example, when differential type edge detection is performed, a differential value between the luminance of the pixel of interest and the luminance of surrounding pixels becomes an index of the pixel of interest. When the index is equal to or greater than a predetermined threshold, the area including the target pixel corresponding to the index is the target area.

In the present embodiment, the creation unit 111 is configured to extract all target regions including pixels whose luminance difference with surrounding pixels is equal to or greater than a predetermined value from the image generated by the first processing unit 110. It is not limited to this configuration. The creation unit 111 may extract at least one target region including a pixel having a luminance difference equal to or greater than a predetermined value from surrounding pixels from the image generated by the first processing unit 110.

In the present embodiment, the image analysis apparatus 101 includes the first processing unit 110, the creation unit 111, the determination unit 112, the determination unit 113, and the second processing unit 114, but is not limited to this configuration. The image analysis apparatus 101 may include some of these components, and a server connected to the image analysis apparatus 101 via a network may include the remaining components. For example, the image analysis apparatus 101 may include the first processing unit 110, and the server may include the creation unit 111, the determination unit 112, the determination unit 113, and the second processing unit 114. In this case, the image analysis apparatus 101 transmits the created image to the server. The server generates a luminance histogram G10 based on the image received from the image analysis apparatus 101. The server determines a luminance threshold based on the generated luminance histogram G10. The server determines whether or not a specific substance (malaria parasite) is present in the image using the luminance threshold. Based on the determination result, the server obtains the number of specific substances present in the image when a specific substance exists in the image. The server transmits the result of the presence / absence of the specific substance and the number of the specific substance, if any, to the image display device 88 via the image analysis apparatus 101.

(Embodiment 2)
This embodiment is different from the first embodiment in that the image analysis apparatus 101 creates a plurality of saliency maps. In the following, the present embodiment will be described focusing on differences from the first embodiment. In addition, the same code | symbol is attached | subjected to the component similar to Embodiment 1, and description is abbreviate | omitted suitably.

The creation unit 111 of the present embodiment creates a plurality of saliency maps using a plurality of extraction processes using the image generated by the first processing unit 110, and brightness based on the created plurality of saliency maps A histogram G10 is created. Here, the plurality of extraction processes include a first extraction process and a second extraction process. The first extraction process is a process of extracting a first candidate region from an image obtained by smoothing the image generated by the first processing unit 110 under a first smoothing condition. The second extraction process is a process of extracting the second candidate region from an image obtained by smoothing the image generated by the first processing unit 110 under the second smoothing condition.

The creation unit 111 applies a Gaussian filter to each pixel of the image generated by the first processing unit 110, and then classifies the image after the Gaussian filter is applied into a plurality of sections according to the first smoothing condition. To divide. Here, the first smoothing condition is to divide the image into areas of m × m pixels. The creation unit 111 creates, for each of the plurality of sections, a first smoothed image that is smoothed by replacing the brightness of the entire section with the average brightness in the section. A first difference image composed of a difference between the image generated by the first processing unit 110 and the first smoothed image is generated. The creation unit 111 extracts all first candidate regions including pixels whose luminance difference with surrounding pixels is greater than or equal to a predetermined value from the first difference image. The process from applying the Gaussian filter to each pixel of the image generated by the first processing unit 110 until extracting the first candidate area corresponds to the first extraction process.

The creation unit 111 applies a Gaussian filter to each pixel of the image generated by the first processing unit 110, and then divides the image after the Gaussian filter is applied into a plurality of sections according to the second smoothing condition. To divide. Here, the second smoothing condition is to divide the image into areas of n × n pixels. Here, n is a value smaller than m. The creation unit 111 creates, for each of the plurality of sections, a second smoothed image that is smoothed by replacing the brightness of the entire section with the average brightness in the section. A second difference image is generated that is the difference between the image generated by the first processing unit 110 and the second smoothed image. The creation unit 111 extracts all second candidate regions including pixels whose luminance difference with surrounding pixels is equal to or greater than a predetermined value from the second difference image. The process from applying the Gaussian filter to each pixel of the image generated by the first processing unit 110 until extracting the second candidate area corresponds to the second extraction process.

The creation unit 111 creates the luminance histogram G10 using all the first candidate areas extracted in the first extraction process and all the second candidate areas extracted in the second extraction process as target areas.

Thereby, a plurality of extraction processes can be performed on the image generated by the first processing unit 110, and a saliency map can be created based on the result.

In the present embodiment, the creation unit 111 is configured to perform the extraction process twice, but is not limited to this configuration. The creation unit 111 may perform the extraction process three times or more.

In the present embodiment, the first smoothing condition and the second smoothing are performed so that the size of the area after the image is divided by the first extraction process is different from the size of the area after the image is divided by the second extraction process. Although the conditions are defined, the smoothing conditions are not limited to this. The smoothing conditions may be set for each extraction process so that the plurality of smoothed images obtained by the plurality of extraction processes have different blurring degrees. For example, the brightness setting method applied to each area after the image is divided by the first extraction process and the brightness setting method applied to each area after the image is divided by the second extraction process are different from each other. You may define 1 smoothing conditions and 2nd smoothing conditions. For example, in the first extraction process, the creation unit 111 replaces the brightness of the entire section with the average brightness in the section for each of the plurality of sections. In the second extraction process, the creation unit 111 replaces the luminance of the entire section with the representative value in the section for each of the plurality of sections. Further, when an image is divided into the same region in each of a plurality of extraction processes (when n = m described above), each extraction process has different division patterns (different division positions) for each extraction process. Different smoothing conditions may be set for.

In the present embodiment, the creation unit 111 creates the luminance histogram G10 using all the first candidate areas obtained by the first extraction process and the second candidate areas obtained by the second extraction process as target areas. Although configured, it is not limited to this configuration. The creation unit 111 creates a luminance histogram G10 using a common area among all the first candidate areas obtained by the first extraction process and the second candidate area obtained by the second extraction process as a target area. May be.

(Modification)
Below, modifications are listed. Note that the modifications described below can be applied in appropriate combination with the above embodiments.

The materials, numerical values, and the like described in the above embodiments are merely preferable examples and are not intended to be limited thereto. Furthermore, in the present invention, the configuration and the shape can be appropriately changed without departing from the scope of the technical idea.

Further, the liquid sample put in the storage space 31 of the filter cartridge 3 may contain a staining solution for staining the nucleic acid of the pathogenic microorganism. In this case, it is not necessary to arrange a fluorescent dye for staining the nucleic acid in the laminated disk body 2. As a staining method using a staining solution, for example, Giemsa staining, acridine orange staining, Wright staining, Jenner staining, Leishmann staining, Romanovsky staining, and the like can be employed. An appropriate staining solution may be used as the staining solution according to the type of pathogenic microorganism and the staining method.

Note that the height of the second well 22 may increase from the inner periphery to the outer periphery of the laminated disc body 2. Thereby, in the disk 1, it is difficult for bubbles to remain in the second well 22. In the disk 1, since the flow rate of the liquid sample decreases from the inner peripheral side to the outer peripheral side of the laminated disk main body 2, the specimen is arranged in the second well 22 so as to be biased toward the outer peripheral side of the laminated disk main body 2. Can be suppressed.

The shape of the laminated disk body 2 viewed from the thickness direction of the laminated disk body 2 is not limited to a circular shape, and may be, for example, an octagonal shape.

Further, a discharge port for discharging the liquid sample may be provided on the outer peripheral portion of the second well 22. The discharge port is, for example, a through hole formed on the upper surface side of the laminated disk main body 2. By providing the discharge port, for example, the liquid sample can be discharged from the discharge port, and the sample adsorbed on the bottom surface of the second well 22 can be separately fixed and stained. Therefore, the disk 1 measured using the detection device 70 can be stored as a specimen sample, for example.

In the disk 1, red blood cells contained in the liquid sample in the first well 21 are moved to the second well 22 by centrifugal force, but the present invention is not limited to this. For example, the red blood cells may be moved from the first well 21 to the second well 22 by generating a pressure difference between the first well 21 and the flow path 23. As one means, a pressure difference can be generated between the first well 21 and the flow path 23 by applying pressure to the first well 21. As a method for applying pressure to the first well 21, pressurization is performed from above the first well 21.

Although the example in which the disc 1, the filter cartridge 3 and the laminated disc main body 2 are used for the examination of red blood cells has been described, the uses of the disc 1, the filter cartridge 3 and the laminated disc main body 2 are not limited to this. For example, the disk 1, the filter cartridge 3, and the laminated disk main body 2 can be used for DNA testing, protein testing, and the like.

(Summary)
As described above, the image analysis system (100) according to the first aspect includes the creation unit (111), the determination unit (112), and the determination unit (113). The creation unit (111) includes a pixel including a pixel whose difference from the luminance of the surrounding pixels from the image (B1) is equal to or greater than a predetermined value and the surrounding pixels, and is formed of an aggregate of pixels smaller than the number of pixels of the image. B20 is extracted as a target area. The creation unit (111) creates a brightness histogram (B10) in which the brightness included in the target region is associated with the frequency at which the brightness appears. The determination unit (112) determines a luminance threshold value from the luminance histogram (G10). The determination unit (113) determines whether or not a specific substance exists in at least the target region extracted by the creation unit (111) using the luminance threshold.

According to this configuration, the image analysis system (100) includes a pixel whose luminance difference from the image (B1) and the surrounding pixels is equal to or greater than a predetermined value and the surrounding pixels, and has fewer pixels than the number of pixels of the image. A luminance histogram (G10) corresponding to the target region made up of Thus, the image analysis system (100) can detect a specific substance from a liquid sample containing a plurality of types of substances even when the brightness of the entire image changes.

In the image analysis system (100) of the second aspect, in the first aspect, the luminance histogram (G10) is a multi-modality including at least the first peak (G11) and the second peak (G12). It is a histogram. The determination unit (112) determines the luminance corresponding to the minimum value of the frequency between the first peak (G11) and the second peak (G12) included in the luminance histogram (G10) as the luminance threshold. .

According to this configuration, the image analysis system (100) divides a luminance histogram (G10) into a region where a specific substance (for example, malaria parasite) exists and a region where the specific substance does not exist, with a luminance threshold as a boundary. Can do.

In the image analysis system (100) of the third aspect, in the first or second aspect, the creation unit (111) differs from the image (B1) with the brightness of surrounding pixels for each of the plurality of extraction processes. Extract at least one candidate region including a pixel having a value equal to or greater than a predetermined value. The creation unit (111) extracts a target region used for creating the luminance histogram (G10) from all candidate regions obtained by a plurality of extraction processes.

According to this configuration, the image analysis system (100) can increase the accuracy of detection of a specific substance.

In the image analysis system (100) of the fourth aspect, in the third aspect, the plurality of extraction processes include a first extraction process and a second extraction process. In the first extraction process, a first candidate region is extracted from an image obtained by smoothing an image under a first smoothing condition. In the second extraction process, a second candidate region is extracted from an image obtained by smoothing the image under the second smoothing condition. The creation unit (111) performs the first extraction process to create a first difference image including a difference between the image (B1) and the first smoothed image smoothed under the first smoothing condition. The creation unit (111) extracts at least one first candidate region including a pixel whose difference from the luminance of surrounding pixels is a predetermined value or more from the first difference image as a candidate region. The creation unit (111) performs the second extraction process, thereby creating a second difference image including a difference between the image (B1) and the second smoothed image smoothed under the second smoothing condition. The creation unit (111) extracts at least one second candidate region including a pixel whose difference from the luminance of surrounding pixels is a predetermined value or more from the second difference image as a candidate region. The creation unit (111) uses all the first candidate areas extracted by the first extraction process and all the second candidate areas extracted by the second extraction process, respectively, for creating the luminance histogram (G10). The target area.

In the image analysis system (100) of the fifth aspect, in the third aspect, the plurality of extraction processes include a first extraction process and a second extraction process. In the first extraction process, a first candidate region is extracted from an image obtained by smoothing an image under a first smoothing condition. In the second extraction process, a second candidate region is extracted from an image obtained by smoothing the image under the second smoothing condition. The creation unit (111) performs the first extraction process to create a first difference image including a difference between the image (B1) and the first smoothed image smoothed under the first smoothing condition. The creation unit (111) extracts at least one first candidate region including a pixel whose difference from the luminance of surrounding pixels is a predetermined value or more from the first difference image as a candidate region. The creation unit (111) performs the second extraction process, thereby creating a second difference image including a difference between the image (B1) and the second smoothed image smoothed under the second smoothing condition. The creation unit (111) extracts at least one second candidate region including a pixel whose difference from the luminance of surrounding pixels is a predetermined value or more from the second difference image as a candidate region. The creation unit (111) determines a common area among all the first candidate areas extracted in the first extraction process and all the second candidate areas extracted in the second extraction process as a luminance histogram (G10). This is the target area used to create

In the image analysis system (100) of the sixth aspect, in any one of the first to fifth aspects, the determination unit (113) uses the luminance threshold value to determine whether a specific substance exists in the image (B1). Determine whether or not.

As a result, the image analysis system 100 determines whether or not a specific substance is present in the entire image. Therefore, the image analysis system 100 detects a specific substance as compared with a case where a part of the image is determined. Accuracy can be improved.

In the image analysis system (100) of the seventh aspect, in any one of the first to sixth aspects, the image (B1) is an image of the sample separated from the liquid sample. The determination unit (113) determines whether or not a specific substance is present in the sample.

According to this configuration, the image analysis system (100) can detect a specific substance from a liquid sample.

In the image analysis system (100) according to the eighth aspect, in any one of the first to seventh aspects, the creation unit (111) determines that the ratio of the number of pixels in the first pixel group to the number of pixels in the target region is an image. The target area is extracted from the image so as to be larger than the ratio of the number of pixels of the second pixel group to the number of pixels of (B1). The first pixel group is a group of pixels in the target area whose difference from the luminance of surrounding pixels is a predetermined value or more. The second pixel group is a group of pixels in the entire image (B1) whose difference from the luminance of surrounding pixels is a predetermined value or more.

The image analysis method according to the ninth aspect includes a creation step, a determination step, and a determination step. The creation step includes a region B20 that includes a pixel whose difference from the luminance of the surrounding pixels from the image (B1) is equal to or greater than a predetermined value and the surrounding pixels, and is composed of a collection of pixels smaller than the number of pixels of the image. Extract as a region. The creating step creates a brightness histogram (G10) in which the brightness included in the target region is associated with the frequency of appearance of the brightness. In the determining step, a luminance threshold is determined from the luminance histogram (G10). In the determination step, it is determined whether or not a specific substance exists in at least the target region extracted in the creation step by using the luminance threshold value.

According to this image analysis method, a specific substance can be detected from a liquid sample containing a plurality of kinds of substances even when the brightness of the entire image changes.

The program according to the tenth aspect is a program for causing a computer to function as the image analysis system (100) according to any one of the first to eighth aspects.

This program can detect a specific substance from a liquid sample containing a plurality of types of substances even when the brightness of the entire image changes.

DESCRIPTION OF SYMBOLS 100 Image analysis system 101 Image analysis apparatus 111 Creation part 112 Determination part 113 Determination part G10 Luminance histogram G11 1st peak part G12 2nd peak part B1 Image (1st image)

Claims

An area including a pixel having a difference from the luminance of surrounding pixels equal to or greater than a predetermined value from the image and the surrounding pixels, and including a collection of pixels smaller than the number of pixels of the image is extracted as the target area, and the target A creation unit that creates a brightness histogram that associates the brightness included in the region with the frequency of appearance of the brightness;
A determination unit for determining a luminance threshold from the luminance histogram;
An image analysis system comprising: a determination unit that determines whether or not a specific substance exists in at least the target region extracted by the creation unit using the luminance threshold.
The luminance histogram is a multimodal histogram including at least a first peak and a second peak.
The determination unit is
The image according to claim 1, wherein a luminance corresponding to the minimum value of the frequency between the first peak and the second peak included in the luminance histogram is determined as the luminance threshold. Analysis system.
For each of a plurality of extraction processes, the creation unit extracts at least one candidate region including a pixel having a difference from the brightness of surrounding pixels equal to or greater than the predetermined value from the image, and is obtained by the plurality of extraction processes. The image analysis system according to claim 1, wherein the target region used for creating the luminance histogram is extracted from all the candidate regions that have been obtained.
The plurality of extraction processes include a first extraction process for extracting a first candidate region from an image obtained by smoothing the image under a first smoothing condition, and a second image obtained by smoothing the image under a second smoothing condition. Including a second extraction process for extracting candidate regions;
The creating unit
By performing the first extraction process, a first difference image made up of a difference between the image and the first smoothed image smoothed under the first smoothing condition is created, and the first difference image Extracting at least one of the first candidate areas including a pixel having a difference from the luminance of the pixel equal to or greater than the predetermined value as the candidate area;
By performing the second extraction process, a second difference image including a difference between the image and the second smoothed image smoothed under the second smoothing condition is created, and the second difference image is Extracting at least one second candidate region including a pixel having a difference from a luminance of the pixel equal to or greater than the predetermined value as the candidate region;
All the first candidate areas extracted in the first extraction process and all the second candidate areas extracted in the second extraction process are set as the target areas used for creating the luminance histogram, respectively. The image analysis system according to claim 3.
The plurality of extraction processes include a first extraction process for extracting a first candidate region from an image obtained by smoothing the image under a first smoothing condition, and a second image obtained by smoothing the image under a second smoothing condition. Including a second extraction process for extracting candidate regions;
The creating unit
By performing the first extraction process, a first difference image made up of a difference between the image and the first smoothed image smoothed under the first smoothing condition is created, and the first difference image Extracting at least one of the first candidate areas including a pixel having a difference from the luminance of the pixel equal to or greater than the predetermined value as the candidate area;
By performing the second extraction process, a second difference image including a difference between the image and the second smoothed image smoothed under the second smoothing condition is created, and the second difference image is Extracting at least one second candidate region including a pixel having a difference from a luminance of the pixel equal to or greater than the predetermined value as the candidate region;
The common area among all the first candidate areas extracted in the first extraction process and all the second candidate areas extracted in the second extraction process is used for creating the luminance histogram. The image analysis system according to claim 3, wherein the image analysis system is a target area.
The determination unit
6. The image analysis system according to claim 1, wherein it is determined whether or not the specific substance is present in the image using the luminance threshold value.
The image is an image of a sample separated from a liquid sample;
The image analysis system according to any one of claims 1 to 6, wherein the determination unit determines whether or not the specific substance is present in the sample.
The creating unit
Extracting the target area from the image such that the ratio of the number of pixels of the first pixel group to the number of pixels of the target area is larger than the ratio of the number of pixels of the second pixel group to the number of pixels of the image;
The first pixel group is a group of pixels in the target area whose difference from the luminance of the surrounding pixels is not less than the predetermined value.
8. The second pixel group is a group of pixels in the entire image having a difference from the luminance of the surrounding pixels equal to or greater than the predetermined value. Image analysis system.
A region including a pixel having a difference from the brightness of surrounding pixels from the image that is equal to or greater than the predetermined value and the surrounding pixels, and including a collection of pixels smaller than the number of pixels of the image as a target region; A creation step of creating a brightness histogram that associates the brightness included in the target region with the frequency of appearance of the brightness;
A determination step of determining a luminance threshold from the luminance histogram;
A determination step of determining whether or not a specific substance is present in at least the target region extracted in the creation step by using the luminance threshold value.
Computer
A program for causing an image analysis system according to any one of claims 1 to 8 to function.