JP2016149998A - Identification apparatus, identification method, and program - Google Patents

Abstract

An object of the present invention is to provide an identification device or the like that can suitably specify a detection target in an image even when conditions such as imaging conditions and illumination conditions fluctuate. An identification apparatus acquires a captured image (microscope image) obtained by observing a microwell accommodated in a petri dish from a microscope, and performs a sharpening process on the acquired microscope image (S102). Then, using the condition data created from the sample image of the positioning unit imaged while changing the focal position (condition), the focal position (condition) of the microscope image is estimated S103, the condition data relating to the estimated focal position (condition) The S104 identification method which detects the positioning part of a microscope image using this and identifies the microwell of the microwell under observation. [Selection] Figure 7

Description

  The present invention relates to an identification device for identifying identification information in an image.

  2. Description of the Related Art Conventionally, images are managed together with identification information by adding identification information to a captured image. For example, Patent Document 1 discloses a culture in which a drop region (cell location) in a culture vessel is detected by image processing, identification information is given to the drop region, and the type of culture vessel and cell identification information are associated and managed. Devices such as container management are disclosed.

JP 2010-220511 A

  By the way, in the example of patent document 1, since a drop area | region is formed by dripping a culture solution on the bottom face of a dish, a convex part is formed with respect to the bottom face of a dish. As described above, when the detection target is formed of a convex portion (or a concave portion), the image characteristics of the detection target may fluctuate greatly depending on conditions such as imaging conditions and illumination conditions, and the detection target may not be detected correctly. .

  The present invention has been made in view of the above-described problems, and an object of the present invention is to suitably detect a detection target in an image even when conditions such as imaging conditions and illumination conditions fluctuate. It is to provide an identification device or the like that can be used.

In order to solve the above-described problem, the first invention includes an acquisition unit that acquires an image including unevenness to be detected, a storage unit that stores a plurality of sample images having different conditions,
Using the target image acquired by the acquisition unit and the sample image stored in the storage unit, an estimation unit for estimating the condition of the target image, and the sample image of the condition estimated by the estimation unit; And a detecting unit that detects the unevenness using the target image.

  The storage unit may store a feature amount of the sample image, and the estimation unit may estimate the condition of the target image based on the feature amount of the sample image and the feature amount of the target image.

  The estimation unit calculates a plurality of feature amounts from the target image and the sample image, and calculates a ratio of the feature amounts that are close to the feature amount of the target image among the plurality of feature amounts of the sample image. Calculating means for selecting, and selecting means for selecting the sample image having a large ratio, and the condition of the selected sample image may be set as the condition of the target image.

  The estimation unit may estimate a condition of the target image based on an image obtained by sharpening the sample image and the target image.

  In order to solve the above-mentioned problem, the second invention is an acquisition step for acquiring an image including irregularities to be detected, a storage step for storing a plurality of sample images under different conditions, and an object acquired by the acquisition step. Using the image and the sample image stored in the storage step, the estimation step for estimating the condition of the target image, the sample image of the condition estimated by the estimation step, and the target image And a detecting step for detecting the unevenness.

  In order to solve the above-mentioned problem, the third invention is characterized in that the computer includes an acquisition unit that acquires an image including unevenness to be detected, a storage unit that stores a plurality of sample images under different conditions, and the acquisition unit. Using the acquired target image and the sample image stored in the storage unit, the estimation unit for estimating the condition of the target image, the sample image of the condition estimated by the estimation unit, and the target image Is a program for functioning as detection means for detecting the unevenness.

  According to the present invention, it is possible to provide an identification device or the like that can suitably detect a detection target in an image even when conditions such as imaging conditions and illumination conditions vary.

System configuration diagram showing a configuration example of a microwell identification system Diagram explaining the arrangement of microwells An enlarged view of the microwell and its vicinity The figure explaining the identification information of a microwell Block diagram showing a configuration example of hardware of the identification device Diagram showing an example of condition data Flow chart showing the flow of microwell identification processing Flow chart showing the flow of estimating condition data close to the focal position The figure which shows an example of the microscope image Im The figure which shows an example of the image data after execution of a sharpening process The figure explaining calculation of the distance of the feature-value of condition data, and the feature-value of the microscope image Im Diagram showing the detection result of the positioning part The figure which shows the detection result (comparative example) of the positioning part

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. The present invention is suitable for the case where the identification device described later automatically identifies microwells, but is not limited thereto. The present invention produces the same effect when applied to the same problems as those described above.

<Microwell identification system 200>
FIG. 1 is a diagram illustrating a configuration example of a microwell identification system 200. The microwell identification system 200 includes an identification device 100, a microscope 101, and a petri dish 1 that is a container in which a plurality of microwells 11 (11a to 11h) are arranged. The microwell identification system 200 observes the microwell 11 held by the petri dish 1 with the microscope 101, and automatically identifies the identification information of the microwell 11 being observed from the captured image.

(Microscope 101)
The microscope 101 is used for observing the microwell 11 in the petri dish 1 placed on the stage 102 and the vicinity thereof. The microscope 101 transmits an image of the object on the stage 102 magnified by the lens to the identification device 100 as a microscope image Im. The identification device 100 acquires the microscope image Im from the microscope 101 and executes a microwell identification process described later. In the present invention, the “target image” means the microscope image Im.

(Microwell 11)
FIG. 2 is a diagram for explaining the arrangement of the microwells 11. As shown in the figure, the petri dish 1 is provided with a cell storage portion 2 in which a plurality of microwells 11 (11a to 11h) having the same size as a cell (fertilized egg in the present embodiment) are arranged. In the present embodiment, an example in which eight microwells 11 are arranged in the cell storage unit 2 is shown. Each microwell 11 is filled with a culture solution and contains a fertilized egg that requires individual management.

  3A and 3B are enlarged views of the microwell 11 and the vicinity thereof. FIG. 3A is a plan view, and FIG. 3B is a cross-sectional view taken along line AA in FIG. In the vicinity of each microwell 11, an identification unit 12 for identifying the microwell 11 and a positioning unit 13 indicating a position serving as a reference when the identification unit 12 is detected on the microscope image Im are provided.

(Identification part 12)
The identification unit 12 is composed of the same circular dots 15, and a maximum of three are provided at the upper, middle and lower positions on the left side of each microwell 11. FIG. 3A shows an example of a pattern of the identification unit 12 that arranges three dots 15 at the upper position, the middle position, and the lower position. The positional relationship among the upper stage position, the middle stage position, and the lower stage position of the identification unit 12 with respect to the corresponding microwell 11 is the same for all the microwells 11 in the petri dish 1.

  Each microwell 11 is assigned unique identification information (hereinafter also referred to as “microwell ID”), and each microwell 11 is based on the number and arrangement combination (pattern) of the dots 15 in the corresponding identification section 12. Can be identified. The identification information will be described later with reference to FIG.

(Positioning part 13)
The shape of the positioning part 13 is the same elliptical shape in all the microwells 11 in the petri dish 1. The positional relationship of the positioning unit 13 with respect to the corresponding microwell 11 and the positional relationship of the identification unit 12 with respect to the positioning unit 13 are the same for all the microwells 11 in the petri dish 1. FIG. 3A shows an example in which the positioning portion 13 is provided at the left position of the microwell 11. The positioning unit 13 has a shape (elliptical shape) different from the dot 15 of the identification unit 12 so that the identification device 100 can distinguish from the identification unit 12.

  FIG. 3B shows a cross-sectional view of FIG. As shown in FIG. 3 (b), the microwell 11 forms a recess suitable for accommodating the fertilized egg 14 individually. Each dot 15 of the identification part 12 forms a circular convex part as shown in the figure. Although not shown, the positioning portion 13 similarly forms an elliptical convex portion. Since the outline of the microwell 11, the identification unit 12, and the positioning unit 13 has an inclination with respect to the bottom surface of the petri dish 1, it becomes dark due to light refraction when observed with the microscope 101, and on the microscope image Im. It becomes possible to recognize them. In addition, each dot 15 of the identification part 12 and the positioning part 13 may form a recessed part unlike the above.

  FIG. 4 is a diagram for explaining the identification information of the microwell 11. 4A to 4H correspond to the microwells 11a to 11h in FIG. 2, and the dots 15 of the identification part 12 of each microwell 11 are represented by black circles. That is, it corresponds to the pattern of dots 15 arranged at the upper position, the middle position, and the lower position of the identification unit 12.

  As shown in the figure, the identification unit 12 is configured by a combination of a different number and arrangement of dots 15 for each corresponding microwell 11. Therefore, the identification device 100 can identify the identification information (microwell ID) of the corresponding microwell 11 by detecting the number and arrangement of the dots 15 in the identification unit 12 from the microscope image Im.

(Identification device 100)
FIG. 5 is a block diagram illustrating a hardware configuration example of the identification device 100. As shown in FIG. 5, the identification device 100 includes, for example, a control unit 21, a storage unit 22, a media input / output unit 23, a communication control unit 24, an input unit 25, a display unit 26, a peripheral device I / F unit 27, and the like. Are connected via a bus 28.

  The control unit 21 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The CPU calls a program stored in the storage unit 22, ROM, storage medium, or the like to a work memory area on the RAM and executes it, drives and controls each device connected via the bus 28, and is performed by the identification device 100. The processing described later is realized. The ROM is a non-volatile memory and permanently holds a computer boot program, a program such as BIOS, data, and the like. The RAM is a volatile memory, and temporarily stores a loaded program, data, and the like, and includes a work area used by the control unit 21 to perform each process.

  The storage unit 22 is an HDD (Hard Disk Drive) or the like, and stores a program executed by the control unit 21, data necessary for program execution, an OS (Operating System), and the like. These program codes are read by the control unit 21 as necessary, transferred to the RAM, and read and executed by the CPU. The storage unit 22 stores in advance information relating to the combination pattern (see FIG. 4) of the dots 15 in the identification unit 12 corresponding to the microwell ID.

The media input / output unit 23 is a media input / output device such as a CD drive, a DVD drive, an MO drive, and a floppy (registered trademark) disk drive, and performs data input / output.
The communication control unit 24 includes a communication control device, a communication port, and the like, and is a communication interface that mediates communication between the computer and the network, and controls communication with other devices via the network.

  The input unit 25 inputs setting values of various parameters, inputs instructions from the user, and has an input device such as a keyboard, a pointing device such as a mouse, and a numeric keypad. The input data is output to the control unit 21.

  The display unit 26 includes, for example, a display device such as a CRT monitor or a liquid crystal panel, and a logic circuit (video adapter or the like) for executing display processing in cooperation with the display device, and is input by the control of the control unit 21. Display information, the acquired microscope image Im, and the like are displayed on the display device.

The peripheral device I / F unit (interface) 27 is a port for connecting the microscope 101 to a computer, and the computer acquires the microscope image Im from the microscope 101 via the peripheral device I / F unit 27. The peripheral device I / F unit 17 is configured by USB, IEEE 1394, RS-232C, or the like, and usually includes a plurality of peripheral devices I / F. The connection form with the peripheral device may be wired or wireless.
The bus 28 is a path that mediates transmission / reception of control signals, data signals, and the like between the devices.

[Condition data 30]
A plurality of condition data 30 having different conditions is stored in advance in the storage unit 22 of the identification device 100. Here, the “condition” refers to an imaging condition, an illumination condition, a microscope setting condition, or the like that affects the image characteristics of the captured image of the microscope 101. The storage unit 22 holds a plurality of condition data 30 having different conditions in association with condition information.

  The gist of the present invention is to estimate the condition of the microscope image Im using a plurality of condition data 30 having different conditions, and to use the condition data 30 related to the estimated condition to determine the positioning unit 13 and the identification unit 12 of the microscope image Im. Is detected with high accuracy. Thereby, for example, even when the conditions such as the imaging conditions, illumination conditions, and microscope setting conditions fluctuate every time the microscope image Im is observed, the positioning unit 13 and the identification unit 12 can be suitably detected. it can.

  In particular, in the present embodiment, the “focus position” of the microscope 101 will be described as an example of the “condition”. The purpose of holding the condition data 30 with different focal positions is as follows. Usually, when the fertilized egg 14 is observed with the microscope 101, the focal position of the microscope 101 is adjusted to the vicinity of the equator of the fertilized egg 14 accommodated in the microwell 11. However, the microwell 11 forms a recess with respect to the horizontal plane of the cell storage unit 2 in the petri dish 1, while the positioning unit 13 and the identification unit 12 form a projection, and thus are accommodated in the microwell 11. When the fertilized egg 14 is observed with the focal position aligned, the focal position is not aligned with the positioning unit 13 and the identification unit 12. In addition, since the size of the fertilized egg 14 changes as a result of growth, the focal position of the microscope 101 varies with each observation.

  For this reason, the contour of the positioning unit 13 is blurred and the degree of blur is not constant, so that the detection accuracy of the positioning unit 13 is reduced, and the detection accuracy of the identification unit 12 that is detected based on the positioning unit 13 is also reduced. Problems arise. Here, if the focus position is a relatively shallow position of the concave portion of the microwell 11, the outline of the positioning unit 13 is clear, and for example, the positioning unit 13 can be detected by edge detection using Robinson Compass Masks. Is possible. However, when the focal position becomes deep in the concave portion of the microwell 11, the focal position moves away from the horizontal plane of the cell storage unit 2, so that the outline of the positioning unit 13 is blurred, and the positioning unit 13 can be detected with high accuracy only by edge detection or the like. It will be difficult to do.

  In this embodiment, in order to obtain a good detection result at any focal position (condition), the data for each different focal position, that is, the positioning unit 13 and the identification unit 12 are photographed while changing the focal position. Condition data 30 created from the sample image is prepared, and the focal position (condition) of the microscope image Im is estimated using the condition data 30. And the positioning part 13 and the identification part 12 are detected with high precision from the microscope image Im using the condition data 30 concerning the estimated focal position (condition). Thereby, erroneous detection due to the influence of the focal position of the microscope image Im can be prevented, and the identification information of the microwell 11 can be well specified.

Here, the condition data 30 of the present embodiment includes the feature amount (feature vector) of each sample image picked up by changing the focal position to N stages (P ₁ , P ₂ ,..., P _N ). . In addition, as the condition data 30, condition data 30A and condition data 30B are prepared. The condition data 30A includes the feature amount of each sample image obtained by imaging the “positioning unit 13” to be detected with the focus position of the microscope 101 changed to N stages, and the focus position (P ₁ , P ₂ ,. .., P _N ) and stored in the storage unit 22. Further, the condition data 30B includes the feature amount of each sample image obtained by imaging the “dot 15 of the identification unit 12” to be detected by changing the focal position of the microscope 101 to N stages, and the focal position (P ₁ , P ₂ ,..., P _N ) and stored in the storage unit 22.

  FIG. 6 is a diagram schematically illustrating the condition data 30 according to the present embodiment. The condition data 30 (condition data 30A, condition data 30B) includes one or a plurality of feature amounts (feature vectors). For example, in the case of FIG. 6, the condition data 30-1 includes three feature amounts (feature vectors) of feature amounts 31a, 31b, and 31c, and the condition data 30-2 includes two feature amounts (features) of feature amounts 32a and 32b. Vector). Since the image characteristics of each sample image are different depending on the focal position, the total number of feature values differs for each condition data 30.

  Note that the feature amount (feature vector) extracted from the sample image (pixel data) is not particularly limited, but in this embodiment, the SURF feature amount is used. The SURF feature amount is a feature point in which a change in shading in an image is detected and pixel values around the feature point are represented by a feature vector. By using the SURF feature value, the contours of the positioning unit 13 and the identification unit 12 in the image can be emphasized, and the contour part can be extracted as a feature point in the image with high accuracy.

  Further, in this embodiment, from the viewpoint of calculation efficiency, the feature amount of the condition data 30 is created in advance and stored in the storage unit 22, but during the microwell identification process, the feature amount is obtained from the sample image each time. You may make it calculate.

<Microwell identification process>
Next, microwell identification processing executed by the identification device 100 will be described with reference to FIGS. FIG. 7 is a flowchart showing the overall flow of the microwell identification process.

  Prior to executing the microwell identification process, the petri dish 1 is placed on the stage 102 of the microscope 101, and the position of the petri dish 1 is adjusted so that the fertilized egg 14 in the microwell 11 becomes the observation target of the microscope 101. It is assumed that the illumination adjustment, focus adjustment, etc. are performed.

(Acquire microscope image)
First, the control unit 21 of the identification apparatus 100 acquires a microscope image Im from the microscope 101 (step S101).
FIG. 9 is a diagram illustrating an example of the acquired microscope image Im. As shown in FIG. 9, the microscopic image Im includes a microwell 11 and a corresponding identification unit 12 (in the example shown in FIG. 9, a pattern in which dots are arranged at the middle position and the lower position, the pattern C in FIG. 4), And the positioning part 13 exists.

(Sharpening)
Next, the control unit 21 of the identification device 100 performs a sharpening process on the microscope image Im for the purpose of increasing the visibility of the microscope image Im (step S102). A method for sharpening an image is not particularly limited, but in this embodiment, CLAHE (Contrast Limited Adaptive Histoequalization) is used. CLAHE can improve the visibility (visibility) of details in an image. Since the image characteristics in the microwell 11 change every moment according to the growth phase of the fertilized egg 14, when global contrast enhancement is performed on the entire image, the visibility of the positioning unit 13 and the identification unit 12 that are detection targets is increased. It may not be able to increase effectively. In this regard, according to CLAHE, since contrast enhancement can be performed for each arbitrary divided region, the visibility of the positioning unit 13 and the identification unit 12 that are detection targets can be reliably increased.

  In addition, the sharpening process of step S102 is an auxiliary process for improving the detection accuracy of the identification unit 12 and the positioning unit 13, and is not necessarily an essential process in the present invention. On the other hand, when the sharpening process in step S102 is performed on the microscope image Im, it should be noted that the feature amount of the condition data 30 needs to be created by performing the same sharpening process on the sample image. .

  FIG. 10 is a diagram illustrating an example of image data obtained by performing the sharpening process in step S102 on the microscope image Im illustrated in FIG. As shown in FIG. 9, it can be confirmed that the contour portions of the positioning portion 13 and the identification portion 12 on the microscope image Im are emphasized by the sharpening process.

(Condition estimation)
Next, a process for estimating the focal position (condition) of the microscope image Im using the condition data 30 will be described with reference to the flowchart of FIG. In the present embodiment, the focal position (condition) of the microscope image Im is estimated using the condition data 30A corresponding to the positioning unit 13.

  First, the control unit 21 of the identification device 100 calculates the distance between the feature amount of the condition data 30A and the feature amount of the microscope image Im (step S201).

  The distance between feature amounts is, for example, a Euclidean distance between feature amounts. That is, if there is a feature quantity X = (x1, x2,..., Xn) having n elements and a feature quantity Y = (y1, y2,..., Yn) having n elements, The distance between the quantities can be calculated as follows.

  The control unit 21 calculates the distance between the feature amounts for all combinations of the feature amounts included in each condition data 30A and the feature amounts calculated from the microscope image Im.

  Next, the control unit 21 of the identification device 100 selects the calculated distance between the feature amounts that is a short distance (step S202). That is, a feature whose feature data 30A is similar to that of the microscope image Im is selected. For example, the control unit 21 selects a feature quantity whose distance between feature quantities is equal to or less than a preset threshold value as a short-distance feature quantity.

  FIG. 11 is a schematic diagram showing selected short-distance feature quantities. In FIG. 11, among the feature amounts included in the condition data 30A-1, the relationship between the feature amount of the microscope image Im and a short distance is indicated by a solid line, and among the feature amounts included in the condition data 30A-2, the microscope image The relationship between the Im feature quantity and the short distance is indicated by a broken line. In the example of FIG. 11, in the condition data 30A-1, the feature quantity 31a ', the feature quantity 31b', and the feature quantity 31c 'are selected as short-distance feature quantities. Further, in the condition data 30A-2, the feature amount 32b 'is selected as a short-distance feature amount.

  The control unit 21 of the identification apparatus 100 then compares the number of feature quantities at a short distance selected in step S202 with respect to the total number of feature quantities included in each condition data 30A (30A-1 to 30A-N). Is calculated (step S203). This ratio indicates how much feature quantity similar to the feature quantity of each condition data 30A (30A-1 to 30A-N) is included in the image of the microscope image Im. It can be determined that the condition data 30A having a larger ratio is closer to the focal position (condition) of the microscope image Im. Hereinafter, the ratio calculated in step S203 is also referred to as “condition matching rate”.

  In the example of FIG. 11, the condition data 30A-1 includes the microscope image Im because all the feature values of the three feature values (31a ′ to 31c ′) are close to one of the feature values of the microscope image Im. The condition coincidence rate of the condition data 30A-1 is “1”. On the other hand, the condition data 30A-2 is the condition data 30A- in the microscope image Im, because only the feature quantity 32b ′ of the two feature quantities (32a ′, 32b ′) is close to the feature quantity of the microscope image Im. The condition matching rate of 2 is “0.5”. That is, in the case of FIG. 11, it can be determined that the condition data 30A-1 is closer to the focal position (condition) of the microscope image Im than the condition data 30A-2.

  The control unit 21 of the identification apparatus 100 determines whether or not the ratio (condition matching rate) in step S203 has been calculated for all the condition data 30A (condition data 30A-1 to 30A-N in FIG. 6). (Step S204) If not calculated (No in Step S204), the process returns to Step S201.

When the ratio (condition matching rate) has been calculated in all the condition data 30A (Yes in step S204), the control unit 21 of the identification device 100 includes the condition data 30A (30A-1 to 30A-N). Condition data 30A having the highest ratio (condition matching rate) is selected, and the focal position (any one of P ₁ , P ₂ ,..., P _N ) associated with the condition data 30A is selected as the focal position of the microscope image Im. (Condition) is determined (step S205). In the following description, in step S205, the control unit 21 selects the condition data 30A-n as the condition data 30A having the highest condition matching rate, and determines the focal position Pn associated with the condition data 30A-n. .

  In the present embodiment, as described above, the condition data 30A having the largest ratio (condition matching rate) is selected as data closest to the focal position (condition) of the microscope image Im. By the way, a method of simply selecting the condition data 30A having the largest number of short-distance feature amounts is also conceivable. However, as described above, since the image characteristics of each sample image are different depending on the focal position, the total number of feature amounts differs between the condition data 30A. Therefore, when the above method is adopted, the condition data 30A close to the focal position (condition) of the microscope image Im is not necessarily selected, and the condition data 30A having a relatively large total number of feature values is selected. It may happen. In this regard, in the present embodiment, as described above, by evaluating with the ratio, data close to the focal position (condition) is surely selected without being influenced by the total number of feature values of the condition data 30. I am doing so.

(Specify microwell ID)
Returning to the flowchart of FIG. In step S104, the control unit 21 of the identification device 100 identifies the microwell ID. Specifically, the control unit 21 first uses the condition data 30A (hereinafter also referred to as “condition matching data”) related to the focal position (condition) estimated in step S205 to determine the positioning unit 13 from the microscope image Im. Is detected.

  For example, the control unit 21 of the identification apparatus 100 calculates the distance between the feature amounts for all combinations of the feature amount of the condition matching data and the feature amount of the microscope image Im, and calculates the coordinates of the feature points related to the feature amount in the short distance. Plot on the microscope image Im. Subsequently, the control unit 21 detects, as the positioning unit 13, a portion in which many plotted feature points are collected and the positional relationship between the plotted feature points and the positional relationship between the feature points of the condition matching data are similar. To do.

  FIG. 12 is a diagram illustrating an example of image data from which the positioning unit 13 has been detected. As shown in FIG. 12, the white dots displayed on the image indicate the coordinates of the feature points whose distance between the feature amounts with the condition matching data in the microscopic image Im is a short distance, and are displayed on the image. A rectangular frame indicating a portion detected as the positioning unit 13.

  Next, when the control unit 21 of the identification device 100 detects the positioning unit 13, a range (in which the identification unit 12 may be displayed on the microscope image Im based on the position and range where the positioning unit 13 exists ( Hereinafter, it is referred to as “identification part existence range”). It is assumed that the positional relationship of the identification unit existence range with respect to the positioning unit 13 is stored in the storage unit 22 in advance. By defining the identification unit existence range, the detection range of the identification unit 12 can be narrowed down.

  Then, the control unit 21 of the identification apparatus 100 refers to the condition data 30B-n of the same focal position Pn (condition) as the focal position Pn (condition) determined in step S205, and determines the feature amount of the condition data 30B-n. Then, the distance between the feature quantities extracted from the image of the identification portion existing range of the microscope image Im is calculated, and the coordinates of the feature points related to the feature quantities at a short distance are plotted on the microscope image Im. Subsequently, the control unit 21 collects a lot of plotted feature points, and identifies a portion in which the positional relationship between the plotted feature points and the positional relationship between the feature points in the condition data 30B-n are similar. This is detected as a dot 15. And the control part 21 specifies applicable microwell ID with reference to the pattern of the dot 15 of the identification part 12 shown in FIG. 4 from the detected number and the detection position of the detected dot 15. FIG.

  As described above, the identification apparatus 100 acquires a captured image (microscopic image Im) obtained by observing the microwell 11 housed in the petri dish 1 from the microscope 101 (step S101), and acquires the acquired microscopic image Im. A sharpening process is performed on the image (step S102). Then, the focal position (condition) of the microscope image Im is estimated using the condition data 30A created from the sample image of the positioning unit 13 captured while changing the focal position (condition) (step S103), and the estimated focal position ( The positioning unit 13 of the microscope image Im is detected using the condition data 30A related to (condition). Further, based on the detection result of the positioning unit 13, the number and arrangement of the dots 15 of the identification unit 12 are detected, and the microwell ID of the microwell being observed is specified (step S104).

  Thereby, in the microscope image Im obtained by observing a plurality of microwells 11 arranged in the petri dish 1 with a microscope, the contours of the positioning unit 13 and the identification unit 12 are blurred, and it is difficult to detect the positioning unit 13 and the identification unit 12. Even if it exists, it becomes possible to specify automatically the microwell ID under observation suitably.

  FIG. 13 is a diagram illustrating a result of detecting the positioning unit 13 by matching the microscope image Im with the condition data 30A (30A-1 to 30A-N) of all the positioning units 13. From the result of FIG. 13, it can be seen that the detection location (black rectangle) is not uniquely determined. That is, if the condition data 30A that does not match the focal position of the microscope image Im is used, the positioning unit 13 cannot be accurately detected. In other words, the positioning unit 13 can be detected satisfactorily by using the condition data 30A that matches the focal position of the microscope image Im as in the present embodiment.

  The preferred embodiments of the identification device 100 and the like according to the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to such examples. For example, the present invention includes various modifications as follows.

<Modification>
As described above, the positioning unit 13 and / or the identification unit 12 that is the detection target may form a concave portion instead of forming a convex portion. Further, the shape, arrangement, and number of the dots 15 of the identification unit 12 shown in FIG. 3 and the like are examples, and various changes can be made. For example, the shape of the dot 15 may be a triangle or a quadrangle, or a plurality of shapes may be used. Also, the shape, arrangement, and number of the positioning portions 13 shown in FIG.

  In this embodiment, the positioning unit 13 is detected using the condition data 30A (condition matching data) having the highest condition matching rate. However, the positioning unit 13 is detected using a plurality of upper condition data 30A having the highest condition matching rate. 13 may be detected. Similarly, the identification unit 12 may be detected using a plurality of condition data 30B.

  In step S104, the identification unit 12 is detected using the condition data 30B having the same focal position as the focal position of the condition data 30A (condition matching data) selected in step S103. Alternatively, the same processing as in step S103 may be performed to select data having a close focal position from the condition data 30B, and the identification unit 12 may be detected using the selected condition data 30B.

  Further, the condition data 30 may include a sample image (pixel data) from which the feature amount is extracted instead of including the feature amount of the image. In this case, the control unit 21 selects the condition data 30 having a focus position close to the microscope image Im by template matching between the microscope image Im and the condition data 30, and detects the positioning unit 13 and the like. Specifically, the control unit 21 scans each sample image (template image) of the “positioning unit 13” captured while changing the focal position over the entire range of the microscope image Im at each scanning position. The similarity between the microscope image Im and the sample image is calculated. In the scanning range, the scanning position having the maximum similarity (hereinafter referred to as “detection position”) and the similarity value are stored in the storage unit 22 for each sample image. The control unit 21 selects a sample image having the maximum similarity among the similarities stored in the storage unit 22 as a sample image (condition matching data) close to the focal position (condition) of the microscope image Im, and stores the storage unit 22. The detected position of the sample image stored in (scanning position with the maximum similarity) is detected as the position where the positioning unit 13 exists. As the similarity, SSD (Sum of Squared Difference), SAD (Sum of Absolute Difference), normalized cross-correlation, or the like can be used. When the positioning unit 13 is detected, the control unit 21 next performs template matching within the identification unit existence range using the template image of the dot 15 of the identification unit 12 to determine the number and arrangement of the dots 15 of the identification unit 12. Detect and identify the microwell ID.

  Further, the estimation of the focal position (condition) in step S103 is performed by comparing the feature amount of the microscope image Im and the feature amount of the condition data 30 as in the present embodiment, and the positioning unit 13 and / or in step S104. The identification unit 12 may be detected by template matching between the microscope image Im and the sample image. In this case, in order to execute the template matching of step S104 with high accuracy, the control unit 21 uses the coordinate transformation matrix obtained in step S103 as a pre-process, and tilts (rotates) the microscope image Im and the sample image. You may perform the image rotation correction | amendment to match. Specifically, from the coordinate correspondence between the feature point coordinates of the condition matching data selected in step S103 and the feature point coordinates of the microscope image Im corresponding to the feature point, the sample image to the microscope image Im (or the microscope image). A coordinate transformation matrix (rotation matrix) from the image Im to the sample image is calculated, and the inclination (rotation) of the microscope image Im and the sample image is matched based on the coordinate transformation matrix. In this case, since the dot 15 on the arc cannot determine the inclination (rotation) of the image, it is desirable to determine the inclination (rotation) of the image by the elliptical positioning unit 13.

  In addition, it is obvious for those skilled in the art that various changes or modifications can be conceived within the scope of the technical idea disclosed in the present application, and these naturally belong to the technical scope of the present invention. Understood

DESCRIPTION OF SYMBOLS 100 ......... Identification apparatus 101 ......... Microscope 102 ......... Stage 200 ......... Microwell identification system 1 ...... Petri dish 2 ......... Cell accommodating part 11 ......... Microwell 12 ......... Identification part 13 ... ...... Positioning part 14 ......... Fertilized egg 15 ......... Dot 21 ......... Control part 22 ......... Storage part 23 ......... Media input / output part 24 ......... Communication control part 25 ......... Input part 26 ... …… Display unit 27 ……… Peripheral device I / F unit 28 ……… Bus 30 ……… Condition data 40 ……… Im feature data Im ……… Microscope image

Claims (6)

An acquisition means for acquiring an image including unevenness to be detected;
Storage means for storing a plurality of sample images having different conditions;
Estimating means for estimating the condition of the target image using the target image acquired by the acquiring means and the sample image stored in the storage means;
Detection means for detecting the unevenness using the sample image of the condition estimated by the estimation means and the target image;
An identification device comprising: The storage means stores a feature amount of the sample image,
The identification apparatus according to claim 1, wherein the estimation unit estimates a condition of the target image based on a feature amount of the sample image and a feature amount of the target image. The estimation means includes
Calculating means for calculating a plurality of feature amounts from the target image and the sample image, and calculating a ratio of the feature amounts that are close to the feature amount of the target image among the plurality of feature amounts of the sample image;
A selection means for selecting the sample image with the large ratio;
Comprising
The identification apparatus according to claim 2, wherein the condition of the selected sample image is set as the condition of the target image.   4. The identification device according to claim 1, wherein the estimation unit estimates a condition of the target image based on a sharpened image of the sample image and the target image. 5. An acquisition step of acquiring an image including unevenness to be detected;
A storage step for storing a plurality of sample images having different conditions;
An estimation step for estimating a condition of the target image using the target image acquired by the acquisition step and the sample image stored in the storage step;
A detection step of detecting the unevenness using the sample image of the condition estimated by the estimation step and the target image;
The identification method characterized by including. Computer
An acquisition means for acquiring an image including unevenness to be detected;
Storage means for storing a plurality of sample images having different conditions;
Estimating means for estimating the condition of the target image using the target image acquired by the acquiring means and the sample image stored in the storage means;
Detection means for detecting the unevenness using the sample image of the condition estimated by the estimation means and the target image;
Program to make it function.