Ophthalmic image processing program, ophthalmic image processing device, and ophthalmic image processing method

Abstract

This invention provides an ophthalmic image processing program, an ophthalmic image processing device, and an ophthalmic image processing method that can appropriately evaluate the condition of the meibomian glands. [Solution] The control unit acquires an eyelid image captured by an ophthalmic imaging device. The image area of ​​the acquired eyelid image includes the eyelid of the eye being examined. The control unit identifies the eyelid region FA that is captured in the image from the image area of ​​the acquired eyelid image. The control unit assigns a region evaluation index 50 to the eyelid region FA in the eyelid image, in a size corresponding to the grade of the condition of the meibomian glands, in order to evaluate the size of the region in which the meibomian glands are present, and then displays the eyelid image on the display unit.

Description

Technical Field

[0001] The present disclosure relates to an ophthalmic image processing program, an ophthalmic image processing apparatus, and an ophthalmic image processing method used for processing eyelid images of an eye to be examined.

Background Art

[0002] A plurality of meibomian glands that secrete oil (lipid), which is one of the components constituting tears, are present in the eyelid. The plurality of meibomian glands pass through the eyelid and have openings slightly inward from the hairline of the eyelashes. The oil secreted from the meibomian glands suppresses excessive evaporation of tear film. When a healthy eyelid is observed, the meibomian glands can be confirmed within a wide area of the eyelid. However, as abnormalities of the meibomian glands progress, the area where the meibomian glands can be identified in the eyelid becomes narrower (in the ophthalmology field, this is sometimes expressed as "the meibomian gland area disappears"). Therefore, by evaluating the state of the meibomian glands in the eyelid, the presence or absence of abnormalities in the meibomian glands can be grasped.

[0003] Methods for evaluating the state of meibomian glands include quantitative evaluation methods and qualitative evaluation methods. For example, the ophthalmic image analysis apparatus described in Patent Document 1 extracts a meibomian gland region from a captured image of an eyelid and obtains distribution information of the meibomian glands in order to perform quantitative evaluation of the meibomian glands. In addition, in the qualitative evaluation method, the state of the meibomian glands is evaluated by comparing the eyelid to be evaluated with a plurality of reference images having different grades of meibomian gland states.

Prior Art Literature

Non-Patent Literature

[0004]

Patent Document 1

Summary of the Invention

Problem to be Solved by the Invention

[0005] Conventional quantitative evaluation methods for meibomian glands suffered from the problem of reduced reproducibility due to changes in imaging conditions when capturing eyelid images. Furthermore, conventional qualitative evaluation methods only compared the eyelid being evaluated with a reference image, leading to discrepancies in evaluation results depending on the evaluator. Therefore, it was difficult to appropriately evaluate the condition of the meibomian glands using conventional techniques.

[0006] A typical object of this disclosure is to provide an ophthalmic image processing program, an ophthalmic image processing device, and an ophthalmic image processing method that can appropriately evaluate the condition of the meibomian glands. [Means for solving the problem]

[0007] An ophthalmic image processing program provided by a typical embodiment of this disclosure is an ophthalmic image processing device that processes an ophthalmic image, which is an image of tissue of an eye to be examined, wherein the ophthalmic image processing program is executed by a control unit of the ophthalmic image processing device, and the ophthalmic image processing device performs the following steps: an image acquisition step of acquiring an eyelid image taken by an ophthalmic image capture device, which includes the eyelid of the eye to be examined within the image region; an eyelid region identification step of identifying the eyelid region from the image region of the acquired eyelid image; and an indexed image display step of applying a region evaluation index for evaluating the size of the region in which meibomian glands exist to the eyelid region in the eyelid image in a size corresponding to the grade of the condition of the meibomian glands, and then displaying the eyelid image on a display unit.

[0008] An ophthalmic image processing device provided in a typical embodiment of this disclosure is an ophthalmic image processing device that processes an ophthalmic image which is an image of the tissue of an eye to be examined, wherein the control unit of the ophthalmic image processing device performs the following steps: an image acquisition step of acquiring an eyelid image which includes the eyelid of the eye to be examined within the image area, taken by an ophthalmic image acquisition device; an eyelid area identification step of identifying the eyelid area from the image area of ​​the acquired eyelid image; and an indexed image display step of applying an area evaluation index for evaluating the size of the area in which meibomian glands exist to the eyelid area in the eyelid image in a size corresponding to the grade of the condition of the meibomian glands, and then displaying the eyelid image on a display unit.

[0009] An ophthalmic image processing method provided by a typical embodiment of this disclosure is an ophthalmic image processing method performed by an ophthalmic image processing device that processes an ophthalmic image which is an image of the tissue of an eye to be examined, and includes: an image acquisition step of acquiring an eyelid image which includes the eyelid of an eye to be examined within the image area, taken by an ophthalmic image acquisition device; an eyelid region identification step of identifying the eyelid region from the image area of ​​the acquired eyelid image; and an indexed image display step of applying a region evaluation index for evaluating the size of the region in which meibomian glands exist to the eyelid region in the eyelid image in a size corresponding to the grade of the condition of the meibomian glands, and then displaying the eyelid image on a display unit.

[0010] According to the ophthalmic image processing program, ophthalmic image processing device, and ophthalmic image processing method relating to this disclosure, the condition of the meibomian glands can be evaluated more appropriately. [Brief explanation of the drawing]

[0011] [Figure 1] This is a block diagram showing the schematic configuration of ophthalmic image processing system 1. [Figure 2] This figure shows an example of an eyelid image 30 to be processed. [Figure 3] This is a flowchart of the ophthalmic image processing performed by the ophthalmic image processing device 10. [Figure 4]This is an explanatory diagram illustrating an example of how to identify the eyelid region (FA) using an interactive foreground extraction algorithm. [Figure 5] This is an explanatory diagram illustrating an example of a method for assigning regional evaluation indices according to the grade of the condition of the meibomian glands. [Modes for carrying out the invention]

[0012] <Overview> The control unit of the ophthalmic image processing device illustrated in this disclosure performs an image acquisition step, an eyelid region identification step, and an indexed image display step. In the image acquisition step, the control unit acquires an eyelid image captured by an ophthalmic imaging device. The image region of the acquired eyelid image includes the eyelid of the eye under examination (in this disclosure, the underside of the eyelid). In the eyelid region identification step, the control unit identifies the eyelid region that is captured in the image from the image region of the acquired eyelid image. In the indexed image display step, the control unit assigns region evaluation indices to the eyelid region in the eyelid image, with sizes corresponding to the grade of the meibomian gland state (e.g., state of meibomian gland disappearance), in order to evaluate the size of the region where meibomian glands are present, and then displays the eyelid image on the display unit.

[0013] According to the technology disclosed herein, within the image area of ​​an eyelid image, the area of ​​the eyelid that is captured in the image is identified, and an area evaluation index is assigned to the identified eyelid area. The area evaluation index is assigned to the eyelid area in a size corresponding to the grade of the condition of the meibomian glands. Therefore, an evaluator evaluating an eyelid image can appropriately evaluate the grade of the condition of the meibomian glands (e.g., the state of disappearance of the meibomian glands) by checking the eyelid area to which the area evaluation index has been assigned. In other words, the grade of the condition of the meibomian glands can be evaluated more accurately than when the eyelid image to be evaluated is simply compared with a reference image. Furthermore, since the evaluator can compare the eyelid area in the eyelid image with the area evaluation index regardless of the shooting conditions when the eyelid image is taken, the decrease in the reproducibility of the evaluation due to changes in shooting conditions is appropriately suppressed. Thus, according to the technology disclosed herein, the condition of the meibomian glands in eyelid images can be appropriately evaluated.

[0014] An ophthalmic image processing device that performs the image acquisition step, the eyelid region identification step, and the indexed image display step can be appropriately selected. For example, an information processing device different from the ophthalmic imaging device (e.g., at least one of a personal computer, tablet terminal, smartphone, and server) may function as the ophthalmic image processing device. In this case, the ophthalmic image processing device may acquire the eyelid image captured by the ophthalmic imaging device via wired communication, wireless communication, or removable memory. Alternatively, the ophthalmic imaging device itself may function as the ophthalmic image processing device.

[0015] The control unit may further perform a grade instruction input step in which the user inputs instructions for selecting a grade that determines the size of the region evaluation index. In the indexed image display step, the control unit may assign region evaluation indexes to the eyelid region that are the size corresponding to the grade selected by the instruction.

[0016] In this case, a region evaluation index having a size corresponding to the grade selected by the user is appropriately applied to the eyelid region in the eyelid image. Therefore, the evaluator can more appropriately evaluate the grade of the meibomian gland state by using the region evaluation index corresponding to the desired grade as a reference.

[0017] When the user designates a grade different from the grade of the region evaluation index applied to the eyelid image being displayed, the control unit may switch the grade of the applied region evaluation index to the newly designated grade. In this case, the evaluator can compare the eyelid region with the region evaluation index while appropriately switching the grade of the region evaluation index. Therefore, the evaluator can more appropriately evaluate the grade of the meibomian gland state.

[0018] However, it is also possible to change the specific method for applying the region evaluation index to the eyelid region. For example, the control unit may cause a plurality of region evaluation indices with different grades to be simultaneously displayed on the eyelid region of one eyelid image. Even in this case, the evaluator can compare the region evaluation index corresponding to each grade with the eyelid region. When applying a plurality of region evaluation indices simultaneously, the control unit may apply each of the plurality of region evaluation indices in a mode that allows the evaluator to distinguish them. Specifically, it is also possible to employ a method of changing the color of each of the plurality of region evaluation indices, a method of attaching a character or the like corresponding to the grade to each of the plurality of region evaluation indices, and the like.

[0019] For each grade of the meibomian gland state, a reference value for the proportion of the disappeared region where the meibomian gland disappears in the eyelid region may be predetermined. In the indexed image display step, the control unit may determine the size of the region evaluation index based on the reference proportion of the disappeared region corresponding to the grade.

[0020] As the grade of the meibomian gland condition worsens, the proportion of the area where the meibomian glands have disappeared tends to increase. Therefore, by determining the size of the area evaluation index based on the proportion of meibomian gland disappearance corresponding to the grade, evaluators can more appropriately assess the grade of the meibomian gland condition using the area evaluation index assigned to the eyelid image.

[0021] Furthermore, specific methods for creating the shape of the region evaluation index can be selected as appropriate. Below, an example of a method for creating the shape of the region evaluation index is described. Generally, the meibomian gland region gradually disappears from one side of the eyelid (upper or lower) to the other as the condition of the meibomian glands worsens. Below, we will explain using the example of the case where the meibomian glands gradually disappear from the lower side of the eyelid to the upper side. Also, the shape of the eyelid is approximately elliptical with the left-right direction as the long axis.

[0022] For example, when the grade of meibomian gland status is the lowest (the meibomian glands are almost not atrophied), the control unit may create a region evaluation indicator in a shape matching the entire eyelid region identified in the eyelid region identification step, and apply the indicator onto the eyelid image. The control unit may also set a horizontally elongated first ellipse that approximates the contour of the eyelid region identified in the eyelid region identification step. When applying a region evaluation indicator corresponding to a grade in which the estimated proportion of the meibomian gland atrophy region is 50%, the control unit may use the long axis of the set first ellipse and the contour of the upper half of the eyelid region identified in the eyelid region identification step as the contour of the region evaluation indicator, and apply the region evaluation indicator onto the eyelid image. The control unit may also set a second ellipse whose long axis matches the long axis of the first ellipse, and whose short axis length is 1 / N (e.g., N=2) of the short axis length of the first ellipse. When applying a region evaluation indicator corresponding to a grade in which the estimated proportion of the meibomian gland atrophy region is 10% to 40% (e.g., 20%), the control unit may use the contour of the lower half of the set second ellipse and the contour of the upper half of the eyelid region identified in the eyelid region identification step as the contour of the region evaluation indicator, and apply the region evaluation indicator onto the eyelid image. Further, when applying a region evaluation indicator corresponding to a grade in which the estimated proportion of the meibomian gland atrophy region is 60% to 90% (e.g., 80%), the control unit may use the contour of the upper half of the set second ellipse and the contour of the upper half of the eyelid region identified in the eyelid region identification step as the contour of the region evaluation indicator, and apply the region evaluation indicator onto the eyelid image.

[0023] In the eyelid region identification step, the control unit may identify the eyelid region by using an interactive foreground extraction algorithm to extract the eyelid region as a foreground. In the interactive foreground extraction algorithm, a foreground is extracted from an image by receiving a user's input instruction for modifying at least one of the foreground and the background in the image, and optimizing foreground estimation for the image in accordance with the input instruction.

[0024] In interactive foreground extraction algorithms, the foreground estimation is optimized in response to user input for modifying at least one of the foreground or background in an image, thereby improving the accuracy of foreground extraction. Therefore, using an interactive foreground extraction algorithm makes it easier to identify the eyelid region in eyelid images with high accuracy.

[0025] As an example, this disclosure uses the GrabCut algorithm, one of the interactive foreground extraction algorithms, to identify the eyelid region from an eyelid image. The GrabCut algorithm uses a Gaussian mixture model (GMM) to describe the pixel distribution and achieves energy minimization through an iterative estimation method. However, it is also possible to identify the eyelid region from an eyelid image using methods other than the GrabCut algorithm.

[0026] However, it is also possible to change the method of identifying the eyelid region from the image area of ​​the eyelid image. For example, a frame of a specific shape or an arbitrary shape may be set within the image area of ​​the eyelid image in response to an operation instruction entered by the user. The area within the set frame may then be identified as the eyelid region.

[0027] The ophthalmic imaging device may perform a notification process to the user to position the eyelids within a predetermined shooting range when capturing eyelid images (for example, displaying a frame indicating the area where the eyelids should be positioned on the display unit). Positioning the eyelids within a predetermined range of the eyelid image is extremely important for improving the accuracy of identifying the eyelid region. Notifying the user to capture eyelid images with the eyelids positioned within the predetermined shooting range appropriately improves the accuracy of identifying the eyelid region.

[0028] <Embodiment> (System Configuration) Hereinafter, one typical embodiment of the present disclosure will be described with reference to the drawings. First, an example of the system configuration of the ophthalmic image processing system 1 in this embodiment will be described in general terms with reference to Figure 1. The ophthalmic image processing system 1 of this embodiment comprises an ophthalmic image processing device 10 and an ophthalmic image acquisition device 20.

[0029] The ophthalmic image processing device 10 processes various data, including ophthalmic images. In this embodiment, the ophthalmic image processing device 10 processes an eyelid image 30 (see Figure 2) that includes the eyelid of the eye being examined (in this embodiment, the underside of the eyelid) within the image area, allowing the user to evaluate the condition of the meibomian glands in the eyelid. In this embodiment, a personal computer (hereinafter referred to as "PC") installed in a medical institution is used as the ophthalmic image processing device 10. However, the ophthalmic image processing device 10 is not limited to a PC. For example, a smartphone or tablet terminal may be used as the ophthalmic image processing device 10. A server may be used as the ophthalmic image processing device 10 separately from or together with the PC. In this case, the server may be, for example, a server of a manufacturer that provides cloud services (a so-called cloud server), or a server other than a cloud server. The ophthalmic image acquisition device 20 itself may function as the ophthalmic image processing device 10. Multiple terminals may cooperate to function as the ophthalmic image processing device 10.

[0030] The ophthalmic image processing device 10 includes a control unit 11 that performs various control processing and a communication interface 14. The control unit 11 includes a CPU 12, which is a controller that manages the control, and a storage device 13 that can store programs and data. The storage device 13 stores an ophthalmic image processing program for executing at least a part of the ophthalmic image processing described later. In this embodiment, the communication interface 14 can connect the ophthalmic image processing device 10 to an external device (for example, an ophthalmic image acquisition device 20, etc.) via a network 5. The ophthalmic image processing device 10 is also connected to an operation unit 16 and a display unit 17.

[0031] The ophthalmic imaging device 20 captures an eyelid image 30 (see Figure 2) that includes the eyelid of the eye being examined (in this embodiment, the underside of the eyelid) within the image area. The ophthalmic imaging device 20 includes a control unit 21 that performs various control processing and a communication I / F 24. The control unit 21 includes a CPU 22, which is a controller that manages the control, and a storage device 23 that can store programs and data. The ophthalmic imaging device 20 also includes an operation unit 26, a display unit 27, and an imaging unit 28. The imaging unit 28 includes various configurations necessary to capture at least the eyelid image 30. In this embodiment, the ophthalmic imaging device 20 captures an infrared image as the eyelid image 30. As shown in Figure 2, the meibomian glands MB present in the eyelid appear in the eyelid image 30 captured by infrared light. The eyelid image 30 is captured with the subject's eyelid turned inside out.

[0032] In this embodiment, the ophthalmic image acquisition device 20 performs a notification process to the user (for example, a display process on the display unit of the frame indicating the area in which the eyelids should be placed) when acquiring an eyelid image 30. As will be described in detail later, in this embodiment, the eyelid region FA (see Figures 4 and 5) within the image area of ​​the eyelid image 30 is identified. It is very important that the eyelids are placed within the predetermined range of the eyelid image 30 in order to improve the accuracy of identifying the eyelid region FA. By notifying the user to acquire the eyelid image 30 with the eyelids placed within the predetermined range, the accuracy of identifying the eyelid region FA is appropriately improved.

[0033] (Ophthalmic image processing) Referring to Figures 2 to 5, an example of ophthalmic image processing performed by the ophthalmic image processing device 10 of this embodiment will be described. The ophthalmic image processing device 10 performs the ophthalmic image processing illustrated in Figure 3 and processes an eyelid image 30 (see Figure 2) that includes the eyelid of the eye being examined (in this embodiment, the underside of the eyelid) within the image area, thereby allowing the user to evaluate the state of the meibomian glands in the eyelid. The ophthalmic image processing is performed by the CPU 12 of the ophthalmic image processing device 10 according to the ophthalmic image processing program stored in the storage device 13.

[0034] As shown in Figure 3, the CPU 12 of the ophthalmic image processing device 10 acquires an eyelid image 30 captured by the ophthalmic image acquisition device 20 (S1). As shown in Figure 2, the image area of ​​the eyelid image 30 includes the eyelid of the eye under examination (in this embodiment, the underside of the eyelid). As mentioned above, the eyelid image 30 in this embodiment is an infrared image in which the meibomian glands MB are captured. Note that in S1, the eyelid image can be acquired by the ophthalmic image processing device 10 via wired communication, wireless communication, or a removable memory, etc.

[0035] Next, the CPU 12 performs histogram equalization and noise reduction processing on the eyelid image 30 acquired in S1, and displays the processed eyelid image 30 on the display unit 17 (S2). Performing at least one of histogram equalization and noise reduction makes it easier to distinguish between the eyelid region and other regions in the eyelid image 30. Figure 2 illustrates an eyelid image 30 in which both histogram equalization and noise reduction have been performed. As an example, in S2 of this embodiment, histogram equalization processing by CLAHE and noise reduction processing by a Median filter are performed on the eyelid image 30. However, it is possible to change the specific methods of histogram equalization and noise reduction. It is also possible to omit at least one of histogram equalization and noise reduction.

[0036] Next, the CPU 12 performs processing to identify the eyelid region from the image region of the eyelid image 30 acquired and processed in S1 to S2 (S5 to S9). In this embodiment, the CPU 12 identifies the eyelid region FA (see Figures 4 and 5) by using an interactive foreground extraction algorithm to extract it from within the image region of the eyelid image 30. In the interactive foreground extraction algorithm, when user instructions are input to modify at least one of the foreground and background in the image, the estimation of the foreground is optimized according to the input, thereby improving the accuracy of foreground extraction. Therefore, by using the interactive foreground extraction algorithm, the eyelid region FA in the eyelid image 30 can be identified with high accuracy. As an example, in this embodiment, the GrabCut algorithm, which is one of the interactive foreground extraction algorithms, is used to identify the eyelid region FA from the eyelid image 30. The GrabCut algorithm uses a Gaussian mixture model (GMM) to describe the pixel distribution and achieves energy minimization through an iterative estimation method.

[0037] The method for identifying the eyelid region FA will be described in detail. The CPU 12 uses the eyelid region as the foreground to be extracted by the interactive foreground extraction algorithm and performs initial foreground extraction on the eyelid image 30 acquired and processed in S1-S2 (S3). In this embodiment, first, the CPU 12 performs initial foreground extraction on the premise that the eyelid is definitely contained within the image region of the eyelid image 30. It is also possible to change the method of initial foreground extraction. For example, the region on which initial foreground extraction is performed may be set within the image region of the eyelid image 30 according to the operation instructions input by the user. Next, the CPU 12 performs a recalculation of foreground extraction on the premise that there is a high probability that the eyelid is located within a predetermined range of a part of the image region of the eyelid image 30 (for example, within a horizontally elongated ellipse in the center of the image region), treating pixels within the predetermined range as foreground pixels and pixels outside the predetermined range as unknown pixels. Through the above processing, initial foreground extraction is performed. The extracted foreground (in this embodiment, the boundary of the eyelid region FA shown in Figure 4) is displayed on the display unit 17.

[0038] Next, the CPU 12 determines whether or not a user instruction has been input to correct the foreground or background that has been extracted at that time (S5). As shown in Figure 4, in this embodiment, the user can input a foreground or background correction instruction to the ophthalmic image processing device 10 by operating the operation unit 16 after understanding the foreground (in Figure 4, the area within the boundary of the eyelid region FA at that time) and background displayed on the display unit 17. For example, Figure 4 Then, a correction area CP, which indicates the area to correct from background to foreground, is input onto the eyelid image 30. Also, Figure 4 Now, a correction area CP for correcting the foreground to the background is input on the eyelid image 30. If no correction instruction has been input from the user (S5:NO), the process proceeds to S8.

[0039] When a correction instruction is input from the user (S5:YES), CPU12 not only corrects the foreground or background according to the input instruction, but also optimizes the foreground estimation according to the input instruction and then performs foreground extraction again (S6). As a result, the foreground is more likely to be extracted with high accuracy. After that, the process moves on to S8.

[0040] The process from S5 to S8 is repeated until an instruction to complete foreground extraction is entered (S8:NO). When the user enters an instruction to complete foreground extraction (S8:YES), the foreground region finally extracted by the interactive foreground extraction algorithm is identified as the eyelid region FA (S9).

[0041] Next, the CPU 12 assigns region evaluation indices 50 (50A, 50B, 50C, 50D) to the corresponding positions of the eyelid region FA on the eyelid image 30 displayed on the display unit 17, in sizes corresponding to the grade of the condition of the meibomian glands MB, for evaluating the size of the region where the meibomian glands MB are located (S10). Therefore, the evaluator evaluating the eyelid image 30 can appropriately evaluate the grade of the condition of the meibomian glands MB (for example, the state of disappearance of the meibomian glands MB) by checking the eyelid region FA to which the region evaluation indices 50 are attached. In other words, the grade of the condition of the meibomian glands MB can be evaluated more accurately than when the eyelid image 30 to be evaluated is simply compared with a reference image. Furthermore, since the evaluator can compare the eyelid region FA in the eyelid image 30 with the region evaluation indices 50 regardless of the shooting conditions when the eyelid image 30 is taken, the decrease in the reproducibility of the evaluation due to changes in shooting conditions is appropriately suppressed. In addition, during the S10 process, which initiates the assignment of region evaluation index 50, the region evaluation index 50 is assigned to the eyelid image 30 with a size corresponding to the default grade.

[0042] Here, with reference to Figure 5, an example of a method for assigning area evaluation index 50 of size according to the grade of the condition of the meibomian glands MB to the eyelid image 30 will be described. Generally, the worse the grade of the condition of the meibomian glands MB, the larger the percentage of the area where the meibomian glands MB disappear (disappearance rate) tends to be. In this embodiment, for each grade of the condition of the meibomian glands MB, a guideline for the percentage of the area where the meibomian glands MB disappear within the eyelid area FA is predetermined. Specifically, in this embodiment, for grade 0, where the condition of the meibomian glands MB is the best, the guideline for the percentage of the area where the meibomian glands MB disappear within the eyelid area FA is set to 0%. For grade 1, the guideline for the percentage of the percentage of the meibomian glands MB disappearing is set to 20%. For grade 2, the guideline for the percentage of the meibomian glands MB disappearing is set to 50%. For grade 3, where the condition of the meibomian glands MB is the worst, the guideline for the percentage of the meibomian glands MB disappearing is set to 80%. In steps S10 and S13 of this embodiment, the CPU 12 determines the size of the region evaluation index 50 based on a guideline for the percentage of disappearance corresponding to the grade. As a result, the evaluator can more appropriately evaluate the grade of the meibomian gland MB (0 to 3 in this embodiment) using the region evaluation index 50 assigned to the eyelid image 30.

[0043] Generally, the meibomian gland region gradually disappears from one eyelid to the other as the condition of the meibomian glands (MB) worsens. Below, we will illustrate the case where the meibomian glands gradually disappear from the lower eyelid to the upper eyelid. If the meibomian glands disappear from the upper eyelid to the lower eyelid, the upper and lower directions described below will be reversed. Also, the shape of the eyelid is roughly elliptical with the left-right direction as the long axis.

[0044] In S10 and S13 of this embodiment, when assigning region evaluation index 50A (see Figure 5) for grade 0, which is the best condition of the meibomian glands MB, the CPU 12 creates the shape of region evaluation index 50A to match the entire eyelid region FA identified in S9 and assigns it to the eyelid image 30. When assigning region evaluation indices 50B, 50C, and 50D (see Figure 5) for other grades, the CPU 12 first sets a horizontally elongated first ellipse that approximates the contour of the eyelid region FA identified in S9. When assigning region evaluation index 50C for grade 2, which is approximately 50% of the meibomian gland disappearance rate, the CPU 12 uses the long axis of the set first ellipse and the contour of the upper half of the eyelid region FA identified in S9 as the contour of region evaluation index 50C and assigns region evaluation index 50C to the eyelid image 30. Furthermore, CPU 12 sets a second ellipse whose major axis coincides with the major axis of the first ellipse, and whose minor axis length is 1 / N of the minor axis of the first ellipse (N=2 in this embodiment). When assigning region evaluation index 50B for Grade 1, where the meibomian gland disappearance rate is estimated to be 10% to 40% (20% in this embodiment), CPU 12 uses the contour of the lower half of the set second ellipse and the contour of the upper half of the eyelid region FA identified in S9 as the contour of region evaluation index 50B and assigns region evaluation index 50B to the eyelid image 30. Furthermore, when assigning region evaluation index 50D for Grade 3, where the meibomian gland disappearance rate is estimated to be 60% to 90% (80% in this embodiment), CPU 12 uses the contour of the upper half of the set second ellipse and the contour of the upper half of the eyelid region FA identified in S9 as the contour of region evaluation index 50D and assigns region evaluation index 50D to the eyelid image 30. As a result, appropriate regional evaluation indices 50A to 50D corresponding to the grade are assigned to the eyelid image 30.

[0045] Returning to the explanation of Figure 3, the CPU 12 determines whether or not a user instruction for selecting a grade has been entered (S12). As an example, in this embodiment, the user can input a grade selection instruction by operating the operation unit 16 and sliding the pointer on the grade selection slider displayed on the display unit 17. If no grade selection instruction has been entered (S12: NO), the process proceeds directly to S14. If a grade selection instruction has been entered (S12: YES), the CPU 12 assigns a region evaluation index 50 of a size corresponding to the grade selected by the input instruction to the eyelid region FA of the eyelid image 30. In other words, if the user has instructed a grade different from the grade of the region evaluation index 50 assigned to the eyelid image 30 during display (S12: YES), the CPU 12 switches the grade of the region evaluation index 50 assigned to the eyelid image 30 to the newly instructed grade. Therefore, the evaluator can more appropriately evaluate the grade of the meibomian gland MB by using the region evaluation index 50 corresponding to the desired grade as a guide. For example, the evaluator can compare the eyelid region FA with the region evaluation index 50 while appropriately switching the grade of the region evaluation index 50. Processes S12 to S14 are repeated until an instruction to terminate the process is entered (S14: NO). When an instruction to terminate the process is entered (S14: YES), the ophthalmic image processing is terminated.

[0046] In this embodiment, once the grade selected by the user is confirmed, the confirmed grade and the eyelid image 30 to which the grade was confirmed are associated and stored in the storage device 13. As a result, the user can then easily and appropriately check the grade of the meibomian gland condition for each subject. The eyelid region FA may also be stored in the storage device 13 along with the eyelid image 30. Furthermore, the stored eyelid region FA may be edited again afterward. In this case, the efficiency of diagnosing the subject's eye using the eyelid image 30 and eyelid region FA is further improved. The storage processing of the grade and eyelid image 30, etc., may be performed automatically or in response to user instructions.

[0047] The technologies disclosed in the above embodiments are merely examples. Therefore, it is possible to modify the technologies exemplified in the above embodiments. First, it is also possible to perform only a part of the multiple technologies exemplified in the above embodiments. For example, in the ophthalmic image processing shown in Figure 3, it is possible to omit either the process of identifying the eyelid region FA (S3-S9) or the process of assigning and displaying region evaluation index 50 with a size corresponding to the grade (S10-S14). [Explanation of symbols]

[0048] 1. Ophthalmic Image Processing System 10. Ophthalmic imaging processing equipment 12 CPU 13 Storage device 17 Display section 20. Ophthalmic imaging equipment 30 Eyelid images 50A~50D Domain Evaluation Indicators FA eyelid area MB Meibomian glands

Claims

1. An ophthalmic image processing program executed by an ophthalmic image processing device that processes ophthalmic images, which are images of tissue of the eye being examined, The ophthalmic image processing program is executed by the control unit of the ophthalmic image processing device, Image acquisition step: Obtaining an eyelid image that includes the eyelid of the eye under examination within the image area, captured by an ophthalmic imaging device, An eyelid region identification step in which the region of the eyelid is identified from the image region of the acquired eyelid image, An indexed image display step involves applying a region evaluation index to the eyelid region in the eyelid image, in a size corresponding to the grade of the condition of the meibomian glands, in order to evaluate the size of the region where the meibomian glands are located, and then displaying the eyelid image on a display unit. An ophthalmic image processing program characterized in that it is executed by the ophthalmic image processing device.

2. An ophthalmic image processing program according to claim 1, The grade instruction input step is further executed, in which the user inputs instructions for selecting the grade that determines the magnitude of the domain evaluation index. An ophthalmic image processing program characterized in that, in the indexed image display step, the region evaluation index of a size corresponding to the grade selected by the instruction is applied to the region of the eyelid.

3. An ophthalmic image processing program according to claim 1 or 2, For each of the aforementioned grades of meibomian gland condition, a predetermined guideline is set for the percentage of the eyelid area where the meibomian glands disappear. An ophthalmic image processing program characterized in that, in the step of displaying an image with indicators, the size of the region evaluation indicator is determined based on a guideline for the proportion of the lost region according to the grade.

4. An ophthalmic image processing program according to any one of claims 1 to 3, An ophthalmic image processing program characterized in that, in the eyelid region identification step, the program takes user instructions to modify at least one of the foreground and background in the image, and extracts the foreground from the image by using an interactive foreground extraction algorithm that extracts the foreground from the image in accordance with the input instructions, thereby identifying the eyelid region as the foreground.

5. An ophthalmic image processing device that processes ophthalmic images, which are images of the tissue of the eye being examined, The control unit of the ophthalmic image processing device is Image acquisition step: Obtaining an eyelid image that includes the eyelid of the eye under examination within the image area, captured by an ophthalmic imaging device, An eyelid region identification step in which the region of the eyelid is identified from the image region of the acquired eyelid image, An indexed image display step involves applying a region evaluation index to the eyelid region in the eyelid image, in a size corresponding to the grade of the condition of the meibomian glands, in order to evaluate the size of the region where the meibomian glands are located, and then displaying the eyelid image on a display unit. An ophthalmic image processing apparatus characterized by performing the following:

6. An ophthalmic image processing method performed by an ophthalmic image processing device that processes ophthalmic images, which are images of tissue of the eye being examined, Image acquisition step: Obtaining an eyelid image that includes the eyelid of the eye under examination within the image area, captured by an ophthalmic imaging device, An eyelid region identification step in which the region of the eyelid is identified from the image region of the acquired eyelid image, An indexed image display step involves applying a region evaluation index to the eyelid region in the eyelid image, in a size corresponding to the grade of the condition of the meibomian glands, in order to evaluate the size of the region where the meibomian glands are located, and then displaying the eyelid image on a display unit. An ophthalmic image processing method characterized by including the following.

Images