Cell identification system and cell identification method

Abstract

To lighten burden of sorting cells.SOLUTION: A cell identification system has an imaging device and an identification device. The imaging device has a well plate, a rotation mechanism, and imaging part. The well plate is provided with a plurality of wells capable of housing cells. The rotation mechanism is provided so as to be capable of rotating the cells housed in the wells. The imaging part is provided so as to be capable of imaging the cells housed in the plurality of wells. The identification device has a rotation control part, an imaging control part, and an identification part. The rotation control part controls the rotation mechanism, and rotates the cells housed in the wells. The imaging control part controls the imaging part each time the cells are rotated by the rotation mechanism, and images the cells housed in the wells. The identification part, by inputting an image of the cells housed in the wells, which is imaged by the imaging part, into a learned model, identifies cells of an excellent condition out of the cells housed in the wells.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Embodiments of the present invention relate to a cell identification system and a cell identification method.

Background Art

[0002] Regenerative medicine using pluripotent stem cells such as iPS cells has already begun to be clinically used. iPS cells are established by giving reprogramming inducing factors to differentiated somatic cells. However, the probability (success rate) of generating cells that can be used as iPS cells is low, and among the cells formed after the introduction of reprogramming inducing factors, the probability of usable cells is approximately 10%.

[0003] Therefore, in the process of establishing and culturing iPS cells, it is necessary to select good iPS cells. Generally, researchers who are good at identification observe cells one by one and identify and select cells in good condition. However, the selection of cells by humans is a heavy burden on researchers and is a factor contributing to the cost of generating iPS cells.

Prior Art Documents

Non-Patent Documents

[0004]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] The problem to be solved by the invention is to reduce the burden of selecting cells.

Means for Solving the Problems

[0006] According to an embodiment, the cell identification system includes an imaging device and an identification device. The imaging device includes a well plate, a rotation mechanism, and an imaging unit. The well plate is provided with a plurality of wells capable of accommodating cells. The rotation mechanism is provided to rotatably support the cells accommodated in the wells. The imaging unit is provided to be able to image the cells accommodated in the plurality of wells. The identification device includes a rotation control unit, an imaging control unit, and an identification unit. The rotation control unit controls the rotation mechanism to rotate the cells accommodated in the wells. The imaging control unit controls the imaging unit to image the cells accommodated in the wells every time the cells are rotated by the rotation mechanism. The identification unit inputs an image of the cells accommodated in the wells, which is imaged by the imaging unit, into a learned model, and identifies healthy cells among the cells accommodated in the wells.

Brief Description of the Drawings

[0007]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

DETAILED DESCRIPTION OF THE INVENTION

[0008] Hereinafter, embodiments will be described with reference to the drawings. Note that, as an example, examples related to iPS cells are shown below, but the cells are not limited to iPS cells and may be appropriately applied to cells other than iPS cells. Also, each of the following embodiments is suitable when the shape of the cells is approximately spherical.

[0009] (First Embodiment) FIG. 1 is a diagram showing an example of the configuration of a cell identification system according to the first embodiment. The cell identification system shown in FIG. 1 includes an imaging device 10 and an identification device 20. In FIG. 1, a schematic cross-sectional view of the imaging device 10 and a block diagram of the identification device 20 are shown. The imaging device 10 and the identification device 20 are connected, for example, by wire. Also, the identification device 20 is connected to a power supply 30 that applies a voltage to the imaging device 10.

[0010] The imaging device 10 is a device that houses iPS cells and can image the housed iPS cells. The imaging device 10 has a well plate 11 for housing iPS cells and a lid portion 12 that covers the well plate 11 so as to seal it.

[0011] FIG. 2 is an example of a top view of the well plate 11 according to the first embodiment. Note that the cross-sectional view shown in FIG. 1 represents the cross-section of the well plate 11 at the two-dot chain line shown in FIG. 2. The well plate 11 has a plurality of recesses called wells formed, for example, in a lattice pattern on the upper surface. This can also be said that a well array is formed on the well plate 11. The well is formed in a cylindrical shape with a substantially circular cross-section. The diameter of the well is about 15 μm to 20 μm, which is larger than the diameter of the iPS cell.

[0012] The well plate 11 is made of, for example, PDMS (polydimethylsiloxane) or Si. Inside the well plate 11, a rotation mechanism for rotating iPS cells selectively accommodated in each well is formed. The rotation mechanism is realized by, for example, a first electrode 111 and a second electrode 112 formed inside the well plate 11. By the first electrode 111 and the second electrode 112, an electrostatically driven torsional oscillator is formed as a rotation actuator for each well, for example.

[0013] Specifically, the first electrode 111 and the second electrode 112 are each formed in a preset direction. For example, in the example shown in FIG. 2, the first electrode 111 is represented by a broken line and is formed along the y-axis direction. The second electrode 112 is represented by a one-dot chain line and is formed along the x-axis direction.

[0014] The first electrode 111 has a movable electrode plate 1111, a fixing part 1112, and a torsion bar 1113. The movable electrode plate 1111 is formed for each well so as to be a partition plate with a cylindrical structure in the well. The movable electrode plate 1111 functions as the bottom of each well. The movable electrode plate 1111 is supported via the torsion bar 1113 by the fixing part 1112 fixed by an insulating layer. A space is formed around the movable electrode plate 1111 so that the movable electrode plate 1111 can rotate about the torsion bar 1113.

[0015] The second electrode 112 has a fixed electrode part 1121. The fixed electrode part 1121 is provided at a position facing the movable electrode plate 1111. The second electrode 112 is formed in an upper layer than the first electrode 111 so that the fixed electrode part 1121 attracts the movable electrode plate 1111 in the vertically upward direction.

[0016] An electrostatically driven torsional oscillator is formed for each well by the movable electrode plate 1111, the fixing part 1112, the torsion bar 1113, and the fixed electrode part 1121.

[0017] A power supply 30 is connected between a first electrode 111 and a second electrode 112 via a switch. When the switch between the first electrode 111 formed in a predetermined column and the second electrode 112 formed in a predetermined row is turned on, a voltage is applied between a movable electrode plate 1111 specified by this row and column and a fixed electrode portion 1121. Thereby, an electrostatic attraction force is generated between the movable electrode plate 1111 and the fixed electrode portion 1121, and the movable electrode plate 1111 rotates about a torsion bar 1113 as an axis.

[0018] Note that the rotation mechanism is not limited to an electrostatically driven torsion oscillator formed by the first electrode 111 and the second electrode 112. For example, other actuators may be used.

[0019] The lid portion 12 shown in FIG. 1 is made of, for example, PDMS or Si. An imaging unit for imaging iPS cells accommodated in the wells of the well plate 11 is provided on a part of the inner upper surface of the lid portion 12. The imaging unit has, for example, a plurality of minute lens elements 1211 and an imaging device 122. Hereinafter, it is assumed that a plurality of wells are formed in a lattice pattern on the well plate 11, and a microlens array 121 in which a plurality of lens elements 1211 are arranged in a lattice pattern in the same manner as the wells.

[0020] The microlens array 121 is an example of an optical system for forming an image of an imaging object on the light receiving surface of the imaging device 122. In the microlens array 121, a plurality of minute lens elements 1211 are arranged on a substrate, for example. The arrangement of the lens elements 1211 coincides with the arrangement of the wells formed in the well plate 11.

[0021] The microlens array 121 is provided with a light source unit 1212. The light source unit 1212 is realized by, for example, a white LED capable of broadband measurement or an LED having a specific wavelength. The light source unit 1212 is provided, for example, between adjacent lens elements 1211 as shown in FIG. 1. Light with energy that does not damage the iPS cells housed in the well is irradiated from the light source unit 1212. Note that the installation location of the light source unit 1212 is not limited to this. When the well plate 11 is formed of, for example, PDMS, the light source unit 1212 may be provided at a position facing the lens element 1211 with the well interposed therebetween.

[0022] The imaging device 122 is an example of an image sensor that converts received light into an electrical signal. The imaging device 122 is realized by, for example, a CCD sensor or a CMOS sensor. The imaging device 122 receives light irradiated from the light source unit 1212 and reflected / scattered by the imaging object through the microlens array 121. The imaging device 122 converts the received light into an image signal as an electrical signal.

[0023] On, for example, the outer upper surface of the lid portion 12, a connection interface 123 for connecting the imaging unit and the identification device 20 is provided. The connection interface 123 has, for example, connection terminals, and by connecting to the identification device 20 through the connection terminals, receives a control signal from the identification device 20 and transmits an image signal output from the imaging device 122 to the identification device 20.

[0024] On the upper surface of the region at one end of the lid portion 12 where the imaging unit is not provided, a first hole 124 is provided. Also, on the upper surface of the region at the other end of the lid portion 12 where the imaging unit is not provided, a second hole 125 is provided. The first hole 124 is used, for example, as an inlet for allowing a solution such as a buffer solution to flow into the imaging device 10. Also, the second hole 125 is used, for example, as an outlet for discharging a solution such as a buffer solution from the imaging device 10. As the buffer solution, for example, PBS (Phosphate Buffered Salts) is used. Note that the diameters of the first hole 124 and the second hole 125 are larger than the diameter of, for example, iPS cells.

[0025] The lid portion 12 has a side wall portion 126. At least a part of the side wall portion 126 is formed with a telescopic portion 1261 that can expand and contract in the extending direction of the side wall portion 126. The telescopic portion 1261 is an example of a distance adjustment mechanism. The telescopic portion 1261 is realized, for example, by a bellows structure, a nested structure, or the like. By the telescopic portion 1261 expanding and contracting, the distance between the upper surface of the well plate 11 and the microlens array 121 changes.

[0026] The well plate 11 and the lid portion 12 are aligned so that the positions of the plurality of wells formed in the well plate 11 and the position of the microlens array 121 provided in the lid portion 12 coincide. When the alignment between the well plate 11 and the lid portion 12 is completed, the lower end portion of the side wall portion 126 of the lid portion 12 is joined to the well plate 11, and the imaging device 10 is created. The height from the lower surface of the well plate 11 to the upper surface of the lid portion 12 is, for example, about 2 mm.

[0027] The identification device 20 shown in FIG. 1 is a device that identifies iPS cells in a good state among the iPS cells housed in the imaging device 10 using a learned model. The identification device 20 includes a processing circuit 21, a memory 22, an input interface 23, a display 24, and a connection interface 25. The processing circuit 21, the memory 22, the input interface 23, the display 24, and the connection interface 25 are communicably connected to each other via a bus, for example.

[0028] The processing circuit 21 is a processor that functions as the center of the identification device 20. The processing circuit 21 realizes the functions corresponding to the program by executing the program stored in the memory 22 and the like.

[0029] The memory 22 is a storage device such as a ROM (Read Only Memory), a RAM (Random Access Memory), an HDD (Hard Disk Drive), an SSD (Solid State Drive), and an integrated circuit storage device that stores various information. Further, the memory 22 may be a driving device or the like that reads and writes various information to and from a portable storage medium such as a CD-ROM drive, a DVD drive, and a flash memory. Note that the memory 22 does not necessarily have to be realized by a single storage device. For example, the memory 22 may be realized by a plurality of storage devices. Further, the memory 22 may be in another computer connected to the identification device 20 via a network.

[0030] The memory 22 stores an identification program and the like according to the present embodiment. Note that these programs may be stored in the memory 22 in advance, for example. Further, for example, they may be stored in a non-transitory storage medium, distributed, read from the non-transitory storage medium, and installed in the memory 22.

[0031] Further, the memory 22 stores, for example, a learned model 221 as an identifier generated by machine learning. The storage of the learned model 221 in the memory 22 may be performed at any time after the manufacture of the identification device 20. For example, it may be at any time during the installation from manufacture to a medical facility or the like, or during maintenance.

[0032] The learned model 221 is obtained by causing a machine learning model to perform machine learning according to a model learning program based on learning data. In the present embodiment, the learned model 221 is, for example, configured to evaluate the state of iPS cells based on the input of an image. In this case, the learning data includes, for example, input data that is an image obtained by imaging iPS cells, and correct output data indicating whether the state of the iPS cells in the image is good or bad.

[0033] FIG. 3 is a diagram for explaining an example of processing when the learned model 221 shown in FIG. 1 is generated based on learning data. The input data is, for example, a plurality of images obtained by imaging a plurality of iPS cells one by one. At this time, for each iPS cell, a plurality of images are used as input data. For example, images captured from a plurality of angles by rotating one iPS cell using a rotation mechanism, and images captured while changing the distance between the iPS cell and the lens are used as input data. Then, for example, whether the iPS cells from which these multiple images were obtained can be cultured as good cells is used as the correct output data. By using as input data an image captured while changing the distance between the iPS cell and the lens, it becomes possible to generate a model that also takes into account the interference pattern (interference fringes) of the light passing through the iPS cells.

[0034] The machine learning model according to this embodiment is, for example, a parameterized composite function in which a plurality of functions are combined. The parameterized composite function is defined by a combination of a plurality of adjustable functions and parameters. The machine learning model according to this embodiment may be any parameterized composite function that satisfies the above requirements, but is assumed to be a multi-layer network model (hereinafter referred to as a multi-layered network). The trained model 221 using the multi-layered network has an input layer for inputting an image, an output layer for outputting whether the state of the iPS cells is good or bad, and at least one intermediate layer or hidden layer provided between the input layer and the output layer. Note that the trained model 221 is assumed to be used as a program module that is part of artificial intelligence software.

[0035] As the multi-layered network according to this embodiment, for example, a deep neural network (DNN), which is a multi-layer neural network targeted for deep learning, is used. As the DNN, for example, a convolution neural network (CNN) targeted for images may be used.

[0036] The input interface 23 shown in FIG. 1 receives various input operations from the user, converts the received input operations into electrical signals, and outputs them to the processing circuit 21. The input interface 23 is connected to input devices such as, for example, a mouse, a keyboard, a trackball, a switch, a button, a joystick, a touch pad, and a touch panel through which an instruction is input by touching an operation surface. Further, the input device connected to the input interface 23 may be an input device provided in another computer connected via a network or the like.

[0037] The display 24 displays various information according to instructions from the processing circuit 21. As the display, for example, any display such as a CRT (Cathode Ray Tube) display, a liquid crystal display, an organic EL display, an LED display, and a plasma display can be appropriately used.

[0038] The connection interface 25 is an interface between the identification device 20 and other devices. Specifically, the connection interface 25 is connected to, for example, the connection interface 123 of the imaging device 10, the power supply 30, and a pump (not shown). The connection interface 25 includes an analog / digital conversion circuit, converts the image signal output from the connection interface 123 into digital-form image data, and outputs the image data to the processing circuit 21 of the identification device 20. Further, the connection interface 25 outputs the control signal generated by the processing circuit 21 to the power supply 30 and the pump.

[0039] The processing circuit 21 realizes functions corresponding to the program by executing an identification program and the like stored in the memory 22. For example, by executing the identification program, the processing circuit 21 has an imaging control function 211, a potential control (voltage control) function 212, a water supply control function 213, an image processing function 214, and an identification function 215. In the present embodiment, a case where the imaging control function 211, the potential control function 212, the water supply control function 213, the image processing function 214, and the identification function 215 are realized by a single processor will be described, but the present invention is not limited thereto. For example, a processing circuit may be configured by combining a plurality of independent processors, and each processor may execute a program to realize the imaging control function 211, the potential control function 212, the water supply control function 213, the image processing function 214, and the identification function 215.

[0040] The imaging control function 211 is a function for controlling the imaging unit provided in the imaging device 10 and is an example of an imaging control unit.

[0041] The potential control function 212 is a function that controls the application of voltage to the electrodes disposed on the well plate 11. By applying voltage to the electrodes, the rotation mechanism provided for each well is driven, and the iPS cells accommodated in the well are rotated. Therefore, the potential control function 212 can also be paraphrased as a rotation control function (rotation control unit) that controls the rotation of the iPS cells accommodated in the well.

[0042] The water supply control function 213 is a function that supplies a solution such as a buffer solution into the imaging device 10 by controlling a pump (not shown). By supplying a solution such as a buffer solution into the imaging device 10, the buffer solution flows over the well, and the iPS cells accommodated in the well are rotated. Therefore, the water supply control function 213 can also be paraphrased as a rotation control function that controls the rotation of the iPS cells accommodated in the well.

[0043] The image processing function 214 is a function that generates an image by performing predetermined image processing on the image data.

[0044] The identification function 215 is a function that uses the learned model 221 with the generated image as input to identify iPS cells in good condition, and is an example of an identification unit.

[0045] Next, the identification process of iPS cells by the identification device 20 configured as described above will be described according to the processing procedure of the processing circuit 21. FIG. 4 is a flowchart showing an example of the operation when the processing circuit 21 shown in FIG. 1 executes the identification process of iPS cells using the learned model 221. In the description of FIG. 4, the case where iPS cells are accommodated in the wells in the imaging device 10 as shown in FIG. 5 will be described as an example. Also, the timing for executing the identification process of iPS cells may be either before or after the differentiation of iPS cells. As shown in FIG. 5, the iPS cells are placed with the movable electrode plate 1111 in the well as the bottom. The inside of the imaging device 10 is filled with, for example, a buffer solution or the like.

[0046] First, when an operator of the identification device 20 inputs an instruction to start the identification process to the identification device 20 via the input interface 23, the processing circuit 21 of the identification device 20 reads out the identification program from the memory 22 and executes the read-out identification program. When the processing circuit 21 executes the identification program, the process shown in FIG. 4 starts.

[0047] In FIG. 4, the processing circuit 21 executes the imaging control function 211. When executing the imaging control function 211, the processing circuit 21 sets the first position of the microlens array 121 with respect to the well plate 11: P = 1, and the first surface of the iPS cells accommodated in the well: S = 1 (step S41). After setting (P = 1, S = 1), the processing circuit 21 acquires image data about the iPS cells (step S42).

[0048] Specifically, the image data about the iPS cells is acquired as follows, for example. The processing circuit 21 turns on the light source unit 1212. FIG. 6 is a diagram showing an example in which the light source unit 1212 is turned on in the imaging device 10 shown in FIG. 5. The white light generated by the light source unit 1212 irradiates the iPS cells accommodated in the well. The irradiated light is reflected and scattered by the surface and internal particles of the iPS cells and reaches the image sensor 122 via the microlens array 121. The image sensor 122 converts the received light into an electrical signal. The electrical signal generated by the image sensor 122 is transmitted as an image signal to the identification device 20 via the connection interface 123. The image signal is converted into image data at the connection interface 25 of the identification device 20 and output to the processing circuit 21.

[0049] When receiving the image data, the processing circuit 21 executes the image processing function 214. When executing the image processing function 214, the processing circuit 21 performs predetermined image processing on the received image data to generate an image. The processing circuit 21 stores the generated image in the memory 22 (step S43).

[0050] FIG. 7 is a diagram showing an example of an image generated by the image processing of the processing circuit 21 according to the first embodiment. As shown in FIG. 7, the image contains iPS cells each housed in a plurality of wells.

[0051] When the image is stored in the memory 22, the processing circuit 21 expands or contracts the telescopic part 1261, which is a distance adjustment mechanism, by the imaging control function 211 (step S44). Thereby, the microlens array 121 is moved to the second position: P = 2. Note that the processing circuit 21 may expand or contract the telescopic part 1261 by a preset distance, for example. When the telescopic part 1261 is expanded or contracted, the processing circuit 21 increments the set position number (step S45). Thereby, (P = 2, S = 1) is set.

[0052] When the position number of the microlens array 121 is incremented, the processing circuit 21 acquires image data about the iPS cells by the imaging control function 211 (step S46). When the image data is acquired, the processing circuit 21 performs predetermined image processing on the acquired image data by the image processing function 214 to generate an image. The processing circuit 21 stores the generated image in the memory 22 (step S47).

[0053] When the image is stored in the memory 22, the processing circuit 21 determines whether imaging has been performed at all preset positions by the imaging control function 211 (step S48). If imaging has not been performed at all positions (No in step S48), the processing circuit 21 shifts the process to step S44 and repeats the processes of steps S44 to S48 until imaging is performed at all preset positions.

[0054] When imaging has been performed at all preset positions (Yes in step S48), the processing circuit 21 determines, by the imaging control function 211, whether imaging of all preset iPS cells has been performed (step S49). When imaging of all iPS cells has not been performed (No in step S49), the processing circuit 21 expands or contracts the expansion and contraction unit 1261 by the imaging control function 211 to return the microlens array 121 to the initial position (the first position) (step S410). When the microlens array 121 is returned to the initial position, the processing circuit 21 sets the first position: P = 1 (step S411). Thereby, (P = 1, S = 1) is set.

[0055] Subsequently, the processing circuit 21 rotates the iPS cells accommodated in the well (step S412). Specifically, the processing circuit 21 executes, for example, the potential control function 212. In the potential control function 212, the processing circuit 21 controls the power supply 30 to apply a voltage between the first electrode 111 and the second electrode 112 of the well plate 11. Thereby, an electrostatic attraction is generated between the movable electrode plate 1111 provided on the first electrode 111 and the fixed electrode portion 1121 provided on the second electrode 112, and the movable electrode plate 1111 rotates about the torsion bar 1113.

[0056] FIG. 8 is a schematic diagram showing an example when the movable electrode plate 1111 rotates. According to FIG. 8, when the movable electrode plate 1111 rotates, the portion functioning as the bottom of the well is inclined in the vertically downward direction. Thereby, the iPS cells placed on the movable electrode plate 1111 roll in the vertically downward direction.

[0057] The processing circuit 21 may apply a voltage between the first electrode 111 and the second electrode 112 multiple times in a pulsed manner, for example. Further, the processing circuit 21 may apply a sinusoidal voltage. By applying a voltage of a plurality of predetermined shapes between the first electrode 111 and the second electrode 112, the rotation and return of the movable electrode plate 1111 are repeated. As a result, the iPS cells placed on the movable electrode plate 1111 are vibrated so as to rotate, and the probability of rotating on the movable electrode plate 1111 is improved. The processing circuit 21 applies a voltage between the first electrode 111 and the second electrode 112, for example, so that the iPS cells rotate about a quarter turn.

[0058] Note that the cross-section of the movable electrode plate 1111 may have a shape along the upper vertical direction as shown in FIG. 9, for example. By devising the shape of the movable electrode plate 1111, it becomes possible to rotate the iPS cells more efficiently.

[0059] Further, when rotating the iPS cells, the processing circuit 21 may also execute a water supply control function 213. In the water supply control function 213, the processing circuit 21 controls a pump to allow a buffer solution to flow in from the first hole 124 of the imaging device 10. As a result, a flow of the buffer solution is generated in the imaging device 10, and a force in the direction from the first hole 124 to the second hole 125 acts on the iPS cells accommodated in the well. Thereby, it becomes possible to rotate the iPS cells more efficiently.

[0060] When all the iPS cells accommodated in the well are rotated about a quarter turn, for example, the processing circuit 21 increments the set image number (step S413). As a result, (P = 1, S = 2) is set.

[0061] When the image number of the iPS cells is incremented, the processing circuit 21 acquires image data about the iPS cells by the imaging control function 211 (step S414). When the image data is acquired, the processing circuit 21 performs predetermined image processing on the acquired image data by the image processing function 214 to generate an image. The processing circuit 21 stores the generated image in the memory 22 (step S415).

[0062] When the processing circuit 21 stores the image in the memory 22, it shifts the process to step S48 and repeats the processes of steps S44 to S48 until imaging at all preset positions is performed. Further, the processing circuit 21 repeats the processes of steps S42 to S415 until imaging of all preset iPS cells is performed.

[0063] When imaging of all preset iPS cells has been performed (Yes in step S49), that is, when images for (P = 1, S = 1), (P = 2, S = 1), …, (P = 1, S = 2), (P = 2, S = 2), …, (P = Pn, S = Sm) are stored in the memory 22, the processing circuit 21 executes the identification function 215. In the identification function 215, the processing circuit 21 uses the learned model 221 stored in the memory 22 to identify iPS cells in good condition among the iPS cells accommodated in the well (step S416).

[0064] FIG. 10 is a diagram for explaining an example of the operation of the identification function 215 shown in FIG. 1. The processing circuit 21 reads out each image captured while changing the vertical position of the microlens array 121 with respect to the well plate 11 and while rotating the iPS cells from the memory 22. That is, the processing circuit 21 reads out images for (P = 1, S = 1), (P = 2, S = 1), …, (P = 1, S = 2), (P = 2, S = 2), …, (P = Pn, S = Sm) from the memory 22.

[0065] The processing circuit 21 extracts the first iPS cells from each of the read images of (P = 1, S = 1), (P = 2, S = 1), …, (P = 1, S = 2), (P = 2, S = 2), …, (P = Pn, S = Sm). According to FIG. 10, the first iPS cells are located, for example, in the topmost row (the first row) and the leftmost column (the first column). The processing circuit 21 inputs the plurality of images extracted for the first iPS cells as input data to the learned model 221. From the learned model 221, information regarding the state of the first iPS cells, for example, whether the state of the first iPS cells is good or bad, is output based on the plurality of images of the input first iPS cells.

[0066] Subsequently, the processing circuit 21 extracts the second iPS cells from each of the images of (P = 1, S = 1), (P = 2, S = 1), …, (P = 1, S = 2), (P = 2, S = 2), …, (P = Pn, S = Sm). According to FIG. 10, the second iPS cells are located, for example, in the second row and the first column. The processing circuit 21 inputs the plurality of images for the second iPS cells as input data to the learned model 221. From the learned model 221, whether the state of the second iPS cells is good or bad is output based on the plurality of images of the input second iPS cells.

[0067] The processing circuit 21 inputs the images for all the iPS cells included in the captured image to the learned model 221 respectively, and causes the learned model 221 to output whether the state of each iPS cell is good or bad. When the states of all the iPS cells included in the captured image are output, the processing circuit 21, for example, displays only the iPS cells with good states on the display 24.

[0068] If there are wells that do not contain iPS cells, the processing circuit 21 extracts the iPS cells contained in the next well from the image after excluding that well.

[0069] iPS cells identified as being in good condition are, for example, collected from the imaging device 10 and transferred to a culture dish in predetermined numbers. The iPS cells transferred to the culture dish form colonies and are cultured in the culture dish as cell colonies.

[0070] As described above, in the first embodiment, the cell identification system includes the imaging device 10 and the identification device 20. The imaging device 10 includes a well plate 11 provided with a plurality of wells capable of accommodating iPS cells, a rotation mechanism 111, 112 provided rotatably with respect to the iPS cells accommodated in the wells, and imaging units 121, 122 provided capable of imaging the iPS cells accommodated in the plurality of wells. Each time the imaging device 10 rotates the iPS cells accommodated in the wells, it images the iPS cells and transmits an image signal to the identification device 20. The identification device 20 stores a learned model in advance. The identification device 20 inputs, as input data, an image generated based on the image signal transmitted from the imaging device 10 into the learned model, and identifies iPS cells in good condition based on the output information. Thereby, it becomes possible to identify iPS cells in good condition without human intervention at the time of creation of iPS cells and in the undifferentiated stage. Further, it becomes possible to identify iPS cells in good condition without human intervention even after the iPS cells have differentiated.

[0071] Also, in the first embodiment, the imaging device 10 further has a distance adjustment mechanism 1261 capable of adjusting the distance between the imaging units 121, 122 and the iPS cells accommodated in the wells. Each time the imaging device 10 changes the distance between the imaging units 121, 122 and the iPS cells accommodated in the wells, it images the iPS cells and transmits an image signal to the identification device 20. As a result, an image that may include an optical interference pattern can be acquired, and it becomes possible to acquire an image related not only to the appearance of the iPS cells but also to the state inside the cytoplasm and the state of the nucleus of the iPS cells. By using an image including an optical interference pattern as input data, the accuracy of identifying iPS cells in good condition is improved.

[0072] In the first embodiment, the imaging unit includes a microlens array 121 and an imaging device 122. As a result, it becomes possible to generate an image including a plurality of iPS cells in a single imaging, and it becomes possible to efficiently image a plurality of iPS cells.

[0073] In the description of the flowchart shown in FIG. 4, the processing starting from the state where iPS cells are accommodated in the wells in the imaging device 10 has been described. A method for efficiently trapping iPS cells in the wells in the imaging device 10 can be realized, for example, as follows.

[0074] FIG. 11 is a diagram showing an example of a top view of a well plate 13 having a function of efficiently trapping iPS cells. Inside the well plate 13, a trapping mechanism for trapping iPS cells in the wells and a rotating mechanism for rotating the iPS cells trapped in the wells are formed. The trapping mechanism and the rotating mechanism are realized, for example, by a third electrode 131, a fourth electrode 132, and a fifth electrode 133 formed inside the well plate 13.

[0075] Specifically, the third electrode 131, the fourth electrode 132, and the fifth electrode 133 are each formed in a preset direction. For example, in the example shown in FIG. 11, the third electrode 131 is represented by a dashed line and is formed in the vicinity of one side surface of the well along the y-axis direction. The fourth electrode 132 and the fifth electrode 133 are represented by a one-dot chain line and are formed so as to sandwich the well along the x-axis direction. The distance between the fourth electrode 132 and the fifth electrode 133 formed with the well in between is shorter than the diameter of the well.

[0076] The third electrode 131 and the fourth electrode 132 are each connected to a power supply 30 via a switch 31 and a switch 32 so that a potential difference can be generated between the third electrode 131 and the fourth electrode 132. Also, the fourth electrode 132 and the fifth electrode 133 are each connected to an AC power supply 40 via a switch 41 and a switch 42 so that a potential difference can be generated between the fourth electrode 132 and the fifth electrode 133.

[0077] FIG. 12 is an enlarged view showing the configuration near the wells of the well plate 13 shown in FIG. 11. FIG. 13 is a cross-sectional view taken along line A-A of the well plate 13 shown in FIG. 12. FIG. 14 is a cross-sectional view taken along line B-B of the well plate 13 shown in FIG. 12.

[0078] The third electrode 131 is disposed above the fourth electrode 132 and the fifth electrode 133. The third electrode 131 has a drive electrode portion 1311 vertically above the region sandwiched between the fourth electrode 132 and the fifth electrode 133. The drive electrode portion 1311 is exposed in a space provided in the well plate 13. The drive electrode portion 1311 is formed to have triangular prism portions 1312 and 1313 with a triangular cross section in the region exposed to the internal space. The triangular prism portions 1312 and 1313 are formed parallel to each other along the y-axis direction. By having the drive electrode portion 1311 with the triangular prism portions 1312 and 1313, it becomes possible to effectively exert an electrostatic force. Note that the drive electrode portion 1311 does not necessarily have to have the triangular prism portions 1312 and 1313.

[0079] The fourth electrode 132 has a movable electrode plate 1321, a torsion bar 1322, and a fixing portion 1323. The movable electrode plate 1321 is formed for each well so as to be a partition plate having a cylindrical structure within the well. The movable electrode plate 1321 functions as the bottom of each well. The movable electrode plate 1321 is supported by the fixing portion 1323 via the conductive torsion bar 1322. A space is formed around the movable electrode plate 1321 so that the movable electrode plate 1321 can rotate about the torsion bar 1322. The movable electrode plate 1321, the torsion bar 1322, and the fixing portion 1323 form an electrostatic rotary actuator as a rotation mechanism for each well.

[0080] The fourth electrode 132 and the fifth electrode 133 function as a trapping mechanism. The distance between the fourth electrode 132 and the fifth electrode 133 formed with the well in between is shorter than the diameter of the well, and the fourth electrode 132 and the fifth electrode 133 are formed to protrude into the well. When an alternating voltage is applied between the fourth electrode 132 and the fifth electrode 133 with the inside of the imaging device 10 filled with a buffer solution, an electrophoretic force is generated in the direction toward the inside of the well as shown in FIG. 14.

[0081] Using the imaging device 10 having the well plate 13 shown in FIGS. 11 to 14, iPS cells are trapped in the wells in the imaging device 10 as follows, for example. For example, the isolated iPS cells are contained in a buffer solution and introduced into the imaging device 10 from the first hole 124 together with the buffer solution.

[0082] FIG. 15 is a diagram showing a connection example of the switches 31, 32, 41, 42 when trapping iPS cells in the wells of the well plate 13. When trapping iPS cells in the wells of the well plate 13, the switches 31, 32 are opened, the switches 41, 42 are closed, and an alternating voltage is applied between the fourth electrode 132 and the fifth electrode 133. The buffer solution containing the iPS cells passes between the well plate 13 in the imaging device 10 and the lid portion 12 and is then discharged from the second hole 125. The iPS cells contained in the buffer solution receive the electrophoretic force generated by the fourth electrode 132 and the fifth electrode 133 as shown in FIG. 16 when the buffer solution flows over the well. Thereby, the iPS cells are trapped in the well.

[0083] FIG. 17 is a diagram showing a connection example of the switches 31, 32, 41, 42 when rotating the iPS cells trapped in the well. When rotating the iPS cells trapped in the well, the switches 31, 32 are closed, the switches 41, 42 are opened, and a voltage is applied between the third electrode 131 and the fourth electrode 132.

[0084] Note that the cells being cultured introduced into the imaging device 10 may be cells in which iPS cells identified as being in good condition have been collected and cultured after undergoing the identification described in the first embodiment. This enables concentration and purification of good iPS cells.

[0085] (Second Embodiment) In the first embodiment, an example of identifying whether the state of iPS cells contained in a well is good or not using the learned model 221 was described. In the second embodiment, an example of identifying whether the state of cell colonies contained in a well is good or not using a learned model will be described.

[0086] FIG. 18 is a diagram showing an example of the configuration of a cell identification system according to the second embodiment. The identification system shown in FIG. 18 includes an imaging device 10A and an identification device 20A. In FIG. 18, a schematic cross-sectional view of the imaging device 10A and a block diagram of the identification device 20A are shown. The imaging device 10A and the identification device 20A are connected, for example, by wire. Further, the identification device 20A is connected to a power supply 30 that applies a voltage to the imaging device 10A.

[0087] The imaging device 10A is a device that houses cell colonies and can image the housed cell colonies. The cell colonies housed in the imaging device 10A are preferably two-dimensional cell colonies in which iPS cells are two-dimensionally distributed. Note that the iPS cells are not necessarily limited to being two-dimensionally distributed. The imaging device 10A has a well plate 11A for housing cell colonies and a lid portion 12A that covers the well plate 11A so as to seal it.

[0088] The well plate 11A is formed of a thermoplastic resin such as PDMS, polyethylene, and polystyrene, which has flexibility, for example. On the upper surface of the well plate 11A, a plurality of recesses called wells are formed, for example, in a lattice pattern. This can also be paraphrased as a well array being formed on the well plate 11A. The wells are formed in a cylindrical shape with a substantially circular cross-section. The diameter of the wells is about several millimeters that can accommodate cell colonies.

[0089] The lid portion 12A is formed of a resin such as silicon like PDMS or ABS, for example. An imaging unit for imaging the cell colonies accommodated in the wells of the well plate 11A is provided on a part of the inner upper surface of the lid portion 12A. The imaging unit has, for example, a plurality of lens elements 1211A and an imaging device 122A. Hereinafter, it is assumed that a plurality of wells are formed in a lattice pattern on the well plate 11A, and a microlens array 121A in which a plurality of lens elements 1211A are arranged in a lattice pattern similar to the wells. The arrangement of the lens elements 1211A in the microlens array 121A coincides with the arrangement of the wells formed in the well plate 11A.

[0090] A light source unit 1212A is provided in the microlens array 121A. The light source unit 1212A is realized by, for example, a white LED or an LED having a specific wavelength. The light source unit 1212A is provided, for example, between adjacent lens elements 1211A as shown in FIG. 18. Note that the light source unit 1212A may be provided on the back side of the well plate 11A. Light with energy that does not damage the cell colonies accommodated in the wells is irradiated from the light source unit 1212A.

[0091] The imaging device 122A is an example of an image sensor that converts received light into an electrical signal. The imaging device 122A is realized by, for example, a CCD sensor, a CMOS sensor, or the like. The imaging device 122A receives light irradiated from the light source unit 1212A and reflected / scattered by the imaging object through the microlens array 121A. When the light source unit 1212A is provided on the back side of the well plate 11A, the imaging device 122A receives light that has passed through the well plate 11A and the imaging object through the microlens array 121A. The imaging device 122A converts the received light into an image signal as an electrical signal.

[0092] On, for example, the outer upper surface of the lid portion 12A, a connection interface 123 for connecting the imaging unit and the identification device 20A is provided. On the upper surface of one end of the lid portion 12A in a region where the imaging unit is not provided, a first hole 124A is provided. Also, on the upper surface of the other end of the lid portion 12A in a region where the imaging unit is not provided, a second hole 125A is provided. The first hole 124A is used, for example, as an inlet for allowing a solution such as a buffer solution to flow into the imaging device 10. The second hole 125A is used, for example, as an outlet for discharging a solution such as a buffer solution from the imaging device 10.

[0093] The lid portion 12A has a side wall portion 126A. At least a part of the side wall portion 126A is formed with, for example, a telescopic portion 1261A that can expand and contract in the extending direction of the side wall portion 126A. The telescopic portion 1261A is an example of a distance adjustment mechanism. The telescopic portion 1261A is realized by, for example, a bellows structure, a nested structure, or the like. By expanding and contracting the telescopic portion 1261A, the distance between the upper surface of the well plate 11A and the microlens array 121A changes.

[0094] The identification device 20A shown in FIG. 18 is a device that identifies cell colonies in a good state among the cell colonies housed in the imaging device 10A using a learned model. The identification device 20A includes a processing circuit 21A, a memory 22A, an input interface 23, a display 24, and a connection interface 25. The processing circuit 21A, the memory 22A, the input interface 23, the display 24, and the connection interface 25 are communicably connected to each other via a bus, for example.

[0095] The processing circuit 21A is a processor that functions as the center of the identification device 20A. The processing circuit 21A realizes the functions corresponding to the program by executing the program stored in the memory 22A and the like.

[0096] The memory 22A is a storage device that stores various information. The memory 22A stores an identification program and the like according to this embodiment. Further, the memory 22A stores, for example, a learned model 221 as an identifier generated by machine learning.

[0097] The processing circuit 21A realizes the functions corresponding to the program by executing the identification program and the like stored in the memory 22A. For example, the processing circuit 21A has an imaging control function 211, a water supply control function 213A, an image processing function 214, an image analysis function 216, and an identification function 215A by executing the identification program. In this embodiment, the case where the imaging control function 211, the water supply control function 213A, the image processing function 214, the image analysis function 216, and the identification function 215A are realized by a single processor will be described, but it is not limited to this. For example, a processing circuit may be configured by combining a plurality of independent processors, and each processor may execute a program to realize the imaging control function 211, the water supply control function 213A, the image processing function 214, the image analysis function 216, and the identification function 215A.

[0098] The water supply control function 213A is a function of supplying a solution such as a buffer solution into the imaging device 10A by controlling a pump (not shown).

[0099] The image analysis function 216 is a function for analyzing the generated image. In the image analysis function 216, the processing circuit 21A extracts, for example, a region containing iPS cells from an image of a cell colony by performing a predetermined contour extraction process on the generated image.

[0100] The identification function 215A is a function for identifying a cell colony in a good state by using the learned model 221 with the image generated by image analysis as an input.

[0101] Next, the identification process of the cell colony by the identification device 20A configured as described above will be described according to the processing procedure of the processing circuit 21A. FIG. 19 is a flowchart showing an example of the operation when the processing circuit 21A shown in FIG. 18 executes the identification process of the cell colony using the learned model 221.

[0102] First, the operator prepares the imaging device 10A. Specifically, the operator places a cell colony in a well formed in the well plate 11A. When the cell colony is placed in the well, the operator aligns the position of the well formed in the well plate 11A with the position of the microlens array 121A provided on the lid portion 12A, and covers the well plate 11A with the lid portion 12A and fixes it.

[0103] When the well plate 11A is covered with the lid portion 12A and fixed, the operator causes the processing circuit 21A to execute the water supply control function 213A via the input interface 23. When the water supply control function 213A is executed, the processing circuit 21A controls the pump and supplies, for example, a buffer solution from the first hole 124A to the imaging device 10A. As a result, the inside of the imaging device 10A is filled with, for example, the buffer solution. FIG. 20 is a diagram showing an example of a cross-sectional view of the imaging device 10A in which a cell colony is accommodated and filled with the buffer solution.

[0104] When the preparation of the imaging device 10A is complete, the operator inputs an instruction to start the identification process into the identification device 20A via the input interface 23. When the start instruction is input, the processing circuit 21A reads the identification program from the memory 22A and executes the read identification program. When the processing circuit 21A executes the identification program, the process shown in FIG. 19 is started.

[0105] In FIG. 19, the processing circuit 21A executes the imaging control function 211. When executing the imaging control function 211, the processing circuit 21A sets the first position: P = 1 of the microlens array 121A with respect to the well plate 11A (step S191). After setting the first position: P = 1, the processing circuit 21A acquires image data based on the image signal of the cell colony obtained by the imaging device 10A (step S192).

[0106] After acquiring the image data, the processing circuit 21A executes the image processing function 214. When executing the image processing function 214, the processing circuit 21A performs predetermined image processing on the received image data to generate an image. The processing circuit 21A stores the generated image in the memory 22A (step S193).

[0107] After storing the image in the memory 22A, the processing circuit 21A expands or contracts the telescopic part 1261A, which is a distance adjustment mechanism, by the imaging control function 211 (step S194). As a result, the microlens array 121A is moved to the second position: P = 2. After expanding or contracting the telescopic part 1261A, the processing circuit 21A increments the set position number (step S195). As a result, the second position: P = 2 is set.

[0108] After incrementing the position number of the microlens array 121A, the processing circuit 21A acquires image data of the cell colony by the imaging control function 211 (step S196). After acquiring the image data, the processing circuit 21A performs predetermined image processing on the acquired image data by the image processing function 214 to generate an image. The processing circuit 21A stores the generated image in the memory 22A (step S197).

[0109] When the processing circuit 21A stores an image in the memory 22A, it determines whether imaging has been performed at all preset positions by the imaging control function 211 (step S198). If imaging has not been performed at all positions (No in step S198), the processing circuit 21A shifts the process to step S194 and repeats the processes of steps S194 to S198 until imaging is performed at all preset positions.

[0110] When imaging has been performed at all preset positions (Yes in step S198), that is, when images for the first position: P = 1, the second position: P = 2,..., the nth position: P = Pn are stored in the memory 22A, the processing circuit 21A executes the image analysis function 216 and the identification function 215A (step S199).

[0111] Specifically, when executing the image analysis function 216 and the identification function 215A, the processing circuit 21A reads out the images for the first position: P = 1, the second position: P = 2,..., the nth position: P = Pn from the memory 22A. The processing circuit 21A extracts the first cell colony from each of the read plurality of images. For example, in the generated image, the first cell colony is located, for example, in the top row (the first row) and the leftmost column (the first column), similar to FIG. 10.

[0112] The processing circuit 21A extracts the region containing iPS cells by performing a predetermined contour extraction process on the image of the extracted first cell colony. FIG. 21 is a schematic diagram showing an example of the image analysis process by the image analysis function 216 shown in FIG. 18. According to FIG. 21, a plurality of regions containing iPS cells are extracted from the image of the cell colony.

[0113] The processing circuit 21A extracts a plurality of regions containing iPS cells by performing contour extraction processing for each of the plurality of images from which the images of the first cell colony have been extracted. At this time, depending on the positional relationship between the well plate 11A and the microlens array 121A, there are images in which the contour of the iPS cells becomes unclear. In such images, the processing circuit 21A may extract the region containing the iPS cells using the positions of the iPS cells extracted from other images among the plurality of images.

[0114] FIG. 22 is a diagram for explaining an example of the operation of the identification function 215A shown in FIG. 18. The processing circuit 21A extracts the region containing the iPS cells from each of the images of the first cell colony at the first position: P = 1, the second position: P = 2, …, the nth position: P = Pn. The processing circuit 21A inputs the plurality of images extracted for the iPS cells as input data to the learned model 221. From the learned model 221, based on the plurality of images of the input iPS cells, whether the state of the iPS cells is good or bad is output.

[0115] The processing circuit 21A determines whether the state of the first cell colony is good or not based on the state of the iPS cells. For example, when the number of outputs indicating that the state is good is equal to or greater than a preset number, the processing circuit 21A determines that the state of the first cell colony is good. On the other hand, when the number of outputs indicating that the state is good is less than the preset number, the processing circuit 21A determines that the state of the first cell colony is bad.

[0116] Subsequently, the processing circuit 21A extracts the second cell colony from each of the images at the first position: P = 1, the second position: P = 2, …, the nth position: P = Pn. The processing circuit 21A repeats the same processing as the processing for the first cell colony to determine whether the state of the second cell colony is good or bad.

[0117] The processing circuit 21A determines the quality of the state for each of all the cell colonies included in the captured image. When determining the state for all the cell colonies included in the captured image, the processing circuit 21A, for example, displays only the cell colonies with a good state on the display 24.

[0118] The cell colonies identified as having a good state are, for example, collected from the imaging device 10A, transferred to a culture dish, and cultured. Then, the cell colonies cultured in the culture dish may be transferred to the imaging device 10A again to have their state determined as good or bad.

[0119] As described above, in the second embodiment, the cell identification system includes the imaging device 10A and the identification device 20A. The imaging device 10A includes a well plate 11A provided with a plurality of wells capable of accommodating cell colonies, imaging units 121A, 122A provided so as to be able to image the cell colonies accommodated in the plurality of wells, and a distance adjustment mechanism 1261A capable of adjusting the distance between the imaging units 121A, 122A and the cell colonies accommodated in the wells. Each time the imaging device 10A changes the distance between the imaging units 121A, 122A and the cell colonies accommodated in the wells, it images the cell colonies and transmits an image signal to the identification device 20A. The identification device 20A stores a learned model in advance. The identification device 20A extracts a region including iPS cells from the image generated based on the image signal transmitted from the imaging device 10A. The identification device 20A inputs the image including iPS cells as input data into the learned model, and based on the output information, identifies the cell colonies with a good state. As a result, an image that may include an optical interference pattern is obtained, and it becomes possible to obtain an image related not only to the appearance of iPS cells but also to the state inside the cytoplasm and the state of the nucleus of iPS cells. Also, at the time of creating iPS cells and in the undifferentiated stage, it becomes possible to identify cell colonies in a good state without human intervention. Furthermore, even after the iPS cells have differentiated, it becomes possible to identify iPS cells in a good state without human intervention.

[0120] In the second embodiment, the imaging unit includes a microlens array 121A and an imaging device 122A. As a result, it becomes possible to generate an image including a plurality of cell colonies in a single imaging, and it becomes possible to efficiently image a plurality of cell colonies.

[0121] In this embodiment, the case where the learned model 221 is generated using, as input data, images of iPS cells captured from a plurality of angles by changing the rotation direction and images captured while changing the distance between the iPS cells and the lens has been described. However, the present invention is not limited to this. The learned model according to this embodiment may be generated using only the images captured while changing the distance between the iPS cells and the lens as input data. That is, in the learned model according to the second embodiment, it is not always necessary to use, as input data, images of iPS cells captured from a plurality of angles.

[0122] In the first and second embodiments, the case where the light source units 1212 and 1212A provided in the imaging devices 10 and 10A are LEDs that generate white light has been described as an example. However, the present invention is not limited to this. The light source units 1212 and 1212A may be laser light sources that generate monochromatic light. When the light source units 1212 and 1212A are realized by monochromatic laser light sources, it is possible to obtain a clear image by staining the iPS cells accommodated in the wells with a fluorescent dye excited by the monochromatic laser light source.

[0123] In addition, in the first and second embodiments, an example of identifying the quality of iPS cells using the learned model 221 was described. However, the present invention is not limited to this. A learned model may be used to identify iPS cells that have reached an appropriate injection timing of a differentiation-inducing factor. At this time, the learned model uses, as input data, a plurality of images captured of iPS cells, including, for example, an image of an iPS cell captured at a timing suitable for injecting a differentiation-inducing factor, and machine learning is performed using whether the iPS cell in the image is at an appropriate timing for injecting a differentiation-inducing factor or not as correct output data and generated. The processing circuit of the identification device inputs an image of an iPS cell obtained by imaging into this learned model to identify an iPS cell that has reached an appropriate injection timing of a differentiation-inducing factor.

[0124] Also, the iPS cell identification process described in the first embodiment and the cell colony identification process described in the second embodiment may be performed in order. That is, for example, the state of a cell colony in which iPS cells identified as being in a good state through the identification described in the first embodiment are collected and cultured may be identified using the method described in the second embodiment. Further, for example, iPS cells may be separated from a cell colony in which a cell colony identified as being in a good state through the identification described in the second embodiment is further cultured, and the state of these iPS cells may be identified using the method described in the first embodiment.

[0125] According to at least one of the embodiments described above, the cell identification system can reduce the burden of selecting cells.

[0126] In the description of the embodiments, the term "processor" means, for example, a CPU (central processing unit), a GPU (Graphics Processing Unit), or a circuit such as an Application Specific Integrated Circuit (ASIC), a Programmable Logic Device (for example, a Simple Programmable Logic Device (SPLD), a Complex Programmable Logic Device (CPLD), and a Field Programmable Gate Array (FPGA)). The processor realizes its functions by reading and executing a program stored in a storage circuit. Note that instead of storing the program in the storage circuit, the program may be directly incorporated into the circuit of the processor. In this case, the processor realizes its functions by reading and executing the program incorporated into the circuit. Each processor in the above embodiments is not limited to being configured as a single circuit for each processor, and a plurality of independent circuits may be combined to form one processor to realize its functions. Further, a plurality of components in the above embodiments may be integrated into one processor to realize its functions.

[0127] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and its equivalent scope.

Explanation of Reference Numerals

[0128] 10, 10A... Imaging device 11, 11A, 13... Well plate 12, 12A… Cover part 20, 20A… Identification device 21, 21A… Processing circuit 22, 22A… Memory 23… Input interface 24… Display 25… Connection interface 30… Power supply 40… AC power supply 31, 32, 41, 42… Switch 111… First electrode 112… Second electrode 131… Third electrode 132… Fourth electrode 133… Fifth electrode 121, 121A… Microlens array 122, 122A… Imaging device 123… Connection interface 124, 124A… First hole 125, 125A… Second hole 126, 126A… Side wall part 211… Imaging control function 212… Potential control function 213, 213A… Water supply control function 214… Image processing function 215, 215A… Identification function 216… Image analysis function 221… Trained model 1111… Movable electrode plate 1112… Fixed part 1113… Twisting rod 1121… Fixed electrode part 1211, 1211A… Lens element 1212, 1212A… Light source part 1261, 1261A… Telescopic part 1311… Driving electrode part 1312, 1313… Triangular prism part 1321… Movable electrode plate 1322… Twisting rod 1323… Fixed part

Claims

1. A well plate provided with a plurality of wells capable of accommodating cells, a rotation mechanism rotatably provided with the cells accommodated in the wells, an imaging unit provided to be able to image the cells accommodated in the plurality of wells and an imaging device having the same; a rotation control unit that controls the rotation mechanism and rotates the cells accommodated in the wells; an imaging control unit that controls the imaging unit and images the cells accommodated in the wells every time the cells are rotated by the rotation mechanism; an identification unit that inputs a plurality of images of the cells accommodated in the wells, imaged by the imaging unit, into a learned model, and identifies the cells in good condition among the cells accommodated in the wells and an identification device having the same and comprising; the imaging device further has a distance adjustment mechanism capable of adjusting the distance between the imaging unit and the cells accommodated in the wells, the imaging control unit controls the imaging unit and images the cells accommodated in the wells every time the distance is changed by the distance adjustment mechanism, the plurality of images are different in the imaging direction with respect to the cells accommodated in the wells and the distance between the imaging unit and the cells accommodated in the wells, a cell identification system.

2. The imaging unit has a microlens array having a plurality of lens elements arranged corresponding to the positions where the plurality of wells are formed, and an imaging element that converts the light that has passed through the microlens array into an electrical signal The cell identification system according to claim 1.

3. The rotation mechanism rotates the cells by vibrating the cells accommodated in the wells from the bottom of the wells. The cell identification system according to claim 1 or 2.

4. The imaging device has a trap mechanism for trapping cells in the wells. The cell identification system according to any one of claims 1 to 3.

5. Image the cells accommodated in a plurality of wells provided in the imaging device with an imaging unit provided in the imaging device, change the distance between the imaging unit and the cells with a distance adjustment mechanism provided in the imaging device, after changing the distance, image the cells with the imaging unit, rotate the cells with a rotation mechanism provided in the imaging device, after rotating the cells, image the cells with the imaging unit, change the distance with the distance adjustment mechanism, after changing the distance, image the cells with the imaging unit, A cell identification method for identifying healthy cells among the cells by inputting a plurality of images with different imaging directions with respect to the cells accommodated in the well and different distances between the imaging unit and the cells accommodated in the well for the cells obtained by the imaging into a learned model.

Images