JP2008079682A - Image analysis apparatus and image analysis program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image analysis apparatus capable of calculating quantitative characteristic values in a blood vessel crossing part from a fundus image and objectively and extremely highly reliably judging the abnormality of a blood vessel on the basis of the calculated results. <P>SOLUTION: The image analysis apparatus calculates the tortuosity factor and the distribution of an aperture of the vein in the crossing part using the data of the vascular wall of the vein extracted around the crossing part. Then, the image analysis apparatus judges the presence/absence of the suspicion of the abnormality of the blood vessel from the calculated results of the tortuosity factor and the distribution of the aperture and outputs the judged result. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image analysis apparatus and an image analysis program. In particular, the present invention relates to an image analysis apparatus and an image analysis program for analyzing a state of a crossing portion of a blood vessel photographed in a fundus image, determining whether there is a suspicion of a blood vessel abnormality, and displaying the determination result.

  It is known that the presence or absence of arteriosclerosis can be determined by analyzing the state of the intersection of the blood vessels imaged in the fundus image. This is because blood vessels in the fundus where arteriosclerotic changes have occurred often have narrow veins at the intersections and bends rapidly.

  In order to analyze the state of the blood vessel portion of the fundus image and determine whether there is a suspicion of abnormality, conventionally, an examiner such as a doctor visually confirms the fundus image one by one and observes the blood vessel portion according to a predetermined criterion. Has been judged whether there is any suspicion of abnormalities. However, visually confirming all the large amounts of fundus images obtained by medical examination etc. has a problem that it takes too much time and is inefficient. In addition, since the determination criteria for arteriosclerotic changes are slightly different for each examiner, there is a risk that the determination results may vary, particularly in an image in which the outline of the blood vessel part is imaged ambiguously.

Patent Document 1 discloses a technique for automatically extracting a contour line from a blood vessel portion imaged unclearly in a fundus image to improve blood vessel analysis accuracy. In Non-Patent Document 1, a portion where a blood vessel is photographed is extracted from a fundus image, and a blood vessel portion near a crossing portion where a plurality of blood vessels cross, and a blood vessel portion separated from the crossing portion by a predetermined distance. Is disclosed, and a technique for determining a blood vessel abnormality based on this value is disclosed.
JP 2002-269539 A JAMIT2006 OP14-6 “Automatic analysis of vascular intersection for supporting diagnosis of hypertension in fundus images” (Presentation at the 25th Annual Meeting of the Japanese Medical Imaging Society, Ryo Takahashi, Yuji Hatanaka, Toshiaki Nakagawa, Yoshinori Hayashi, Satoshi Fujita and others )

  The technique disclosed in Non-Patent Document 1 is a technique that makes it possible to quantitatively and efficiently determine whether there is a suspicion of an abnormality in a blood vessel. However, in the technique of Non-Patent Document 1, since the caliber ratio is calculated from the two calibers of the blood vessel part, there is a fear that the change in the caliber of the part that was not used for the caliber ratio may be overlooked. Further, when the contrast of the fundus image is low, the error in the values of the two calibers becomes large, and the error in the caliber ratio increases accordingly, which may impair the reliability of the determination result. .

  On the other hand, the venous flexion degree at the intersection, which is as important as the change in the diameter of the blood vessel in determining the abnormality of the blood vessel, has hardly been quantified so far. In fundus images, the veins at the intersection are often captured in a state hidden by the other blood vessel that intersects, and the contrast at the intersection is low overall. However, this is because it is difficult to accurately calculate the degree of bending of the veins at the intersection.

  Conventionally, visual inspection has been most widely performed in order to determine the state of a blood vessel portion photographed in a fundus image. In order to make a quick and reliable determination, a technique has been disclosed in which a caliber ratio in the vicinity of a blood vessel intersection is calculated and used for the determination. However, there is a need for a technique that can determine the state of blood vessels more quickly and with high reliability.

  The present invention has been made to solve the above-described problems, and calculates the degree of venous flexion and the diameter of the vein at the intersection, and uses this calculation result for the determination of an abnormality in the blood vessel. It is an object of the present invention to provide a technique capable of objectively and extremely reliably determining abnormality based on quantitative characteristics.

  The invention of claim 1 relates to an image analysis apparatus for analyzing a crossing portion of a blood vessel captured in an image obtained by capturing a fundus of a subject. The image analysis apparatus according to the present invention determines a crossing region determining unit that determines a crossing region including a crossing portion of a blood vessel in a fundus image, and determines whether a blood vessel included in the crossing region is an artery or a vein. The distribution of arteriovenous determination means, the blood vessel wall extracting means for extracting the blood vessel wall of the blood vessel included in the crossing region, and the distribution of the diameter of the vein in the crossing region and the bending degree from the extracted data of the position of the blood vessel wall of the vein And a vein analysis means for calculating. The vein analysis means of the image analysis apparatus of the present invention calculates the position of the blood vessel wall of the vein at the crossing portion based on the data of the position of the blood vessel wall extracted from the crossing region, and distributes and curves the caliber of the vein diameter at the crossing portion. The degree is calculated.

  The image analysis apparatus of the present invention extracts a blood vessel wall from a blood vessel extending in the crossing region. Then, based on the data of the position of the vein in the vein, the position of the blood vessel wall of the vein hidden in the intersection is estimated. From the data of the position of the blood vessel wall of the blood vessel in the crossover region, including the estimated position of the blood vessel wall of the crossing portion, the distribution of the diameter of the vein and the degree of bending can be calculated.

  An image analysis apparatus according to a second aspect of the present invention includes a vascular state determination unit that determines whether or not there is a suspicion of a blood vessel abnormality in the fundus from the calculation result of the distribution of the caliber of the veins included in the crossover region and the degree of bending. The apparatus further comprises display means for displaying the judgment result of the means.

  Abnormal veins in the fundus are often narrowed at the crossovers and bent rapidly. The image analysis apparatus of the present invention calculates the distribution of the caliber and the degree of bending of the veins included in the crossing region, determines whether or not the calculation result shows an abnormal characteristic, and displays the determination result Can do.

  A third aspect of the present invention relates to an image analysis program for analyzing a cross section of a blood vessel of a fundus from data of an image obtained by photographing a fundus of a subject. The image analysis program of the present invention determines a crossover region including a crossing portion of a blood vessel in a fundus image, determines whether a blood vessel included in the crossing region is an artery or a vein, and includes the crossing region. The blood vessel wall of the extracted blood vessel is extracted, the data of the position of the blood vessel wall of the vein in the extracted blood vessel is calculated, and the data of the position of the blood vessel wall of the vein at the intersection is calculated from the data of the calculated blood vessel wall position And calculating the distribution of the caliber and the degree of curvature of the veins in the crossing region based on the calculated position data.

  The image analysis apparatus and the image analysis program according to the present invention make it possible to calculate the distribution of the caliber and the degree of bending of the vein at the intersection of blood vessels, which has been difficult to calculate from the fundus image.

  Furthermore, based on the distribution of the caliber of the vein at the intersection of the blood vessels calculated by the image analysis apparatus of the present invention and the value of the degree of flexion, the presence or absence of a suspected vascular abnormality is determined with high reliability, and the determination result is displayed. Became possible.

  Hereinafter, an embodiment in which a blood vessel of a fundus image is analyzed by the image analysis apparatus of the present invention will be described in detail with reference to the drawings.

  (Embodiment 1) FIG. 1 shows a block diagram of an image analyzing apparatus 1 in this embodiment. The image analysis apparatus 1 is composed of a single computer. The image analysis apparatus 1 includes a blood vessel extraction processing unit 2, a crossing region determination processing unit 4, an arteriovenous determination processing unit 6, a blood vessel wall extraction processing unit 8, a vein analysis processing unit 10, and a blood vessel state determination processing unit 12. And a display processing unit 16. These processing units are stored in a format that can be executed in the storage unit of the computer, and the analysis of the fundus image proceeds by being executed by the CPU of the computer.

  Furthermore, the image analysis device 1 includes an input unit 14 that can input a fundus image, and an output unit 18 that can display the input fundus image and the result of the analysis performed by the image analysis device 1. ing.

  An example of the fundus image analyzed by the image analysis apparatus 1 of the present embodiment is shown in FIG. The fundus image analyzed in the present embodiment is a digital color image, and includes an optic nerve head 21, a retina 22, and a plurality of blood vessels 24 extending from the optic nerve head 21 and running above the retina 22. Have been filmed. The image analysis apparatus 1 analyzes the veins in the blood vessel 24 of the fundus image 20 in particular, and determines whether there is an abnormality.

  FIG. 3 shows a flow of analysis performed on the fundus image 20 by the image analysis apparatus 1. Hereinafter, detailed analysis contents will be described along the flow of FIG. The input unit 14 of the image analysis apparatus 1 inputs pixel data of the fundus image 20 in step S2. The pixel data of the fundus image 20 in the present embodiment includes R (red), G (green), and B (blue) data for each pixel.

  The blood vessel extraction processing unit 2 of the image analysis device 1 identifies the blood vessel 24 from the fundus image 20 in step S4. The blood vessel 24 is identified by binarizing the input pixel value data of the fundus image 20. Binarization is performed by identifying a pixel having a pixel value unique to the blood vessel part and a pixel having a pixel value other than the blood vessel part, and assigning 0 or 1 data to each identified pixel. . Furthermore, the blood vessel extraction processing unit 2 performs a thinning process on the identified blood vessel 24 to convert the blood vessel 24 into a line segment having a width of 1 pixel. Then, the position information of the thin line is temporarily stored as the position information of the blood vessel 24.

  In step S <b> 6, the cross region determination processing unit 4 of the image analysis apparatus 1 detects the cross portion 26 of the blood vessel 24 and sets the cross region 28 for the fundus image 20 on which the blood vessel extraction processing has been performed. The crossing area determination processing unit 4 uses the information on the position of the blood vessel 24 to detect a crossing portion 26 where the two blood vessels 24 cross each other. Then, a square crossing region 28 having a predetermined area is set so that the crossing portion 26 is at the center. When a large number of crossing portions 26 are detected, the crossing region determination processing unit 4 refers to the binarized data obtained by processing the fundus image 20 by the blood vessel extraction processing unit 2 and connects the thicker blood vessels 24 to each other. 1 or two or more crossing regions 28 can be set by detecting the crossing portion 26 where the crossing points.

  In step S8, the arteriovenous determination processing unit 6 of the image analysis apparatus 1 analyzes the pixel value of the crossing region 28 set by the crossing region determination processing unit 4 in detail, and the blood vessel 24 included in the crossing region 28 becomes an artery. Or whether it is a vein. Since the chromaticity is different between the artery and the vein, it is possible to determine whether the intersecting blood vessel is an artery or a vein by detailed comparison of the R, G, and B values. As a result of the determination, the arteriovenous determination processing unit 6 in this embodiment cancels the region setting for the crossing region 28 in which it is clear that the same type of blood vessel is crossing in the crossing unit 26. For the crossing region 28 determined to intersect the artery and the vein, the information on the position of the blood vessel 24 is associated with the information on the blood vessel type indicating whether the blood vessel 24 is an artery or a vein, Temporarily store again.

  In step S <b> 10, the blood vessel wall extraction processing unit 8 of the image analysis apparatus 1 extracts arterial and venous blood vessel walls included in the crossover region 28. As shown in FIG. 2, the blood vessel 24 has a lower brightness value calculated based on the R, G, and B values than the region of the retina 22 other than the blood vessel 24. In addition, in the portion that becomes the boundary between the retina 22 and the blood vessel 24, the amount of change in the luminance value is large, and the boundary is clear. Therefore, by evaluating pixel data for each pixel in a region with a large change in luminance value, selecting a pixel having a luminance equal to or lower than a predetermined threshold value, and connecting consecutive points of this pixel, the blood vessel 24 The wall can be determined.

  FIG. 4 shows an example of a blood vessel wall extraction result performed by the blood vessel wall extraction processing unit 28 for the artery 30 and the vein 32 determined to be included in the crossing region 28. Depending on the change in luminance and the threshold value used by the blood vessel wall extraction processing unit 8, the blood vessel wall in the low contrast portion of the crossing portion is difficult to be extracted. For this reason, for the vein 32 of the present embodiment, two pairs of blood vessel walls 33, 34, 35, and 36 that are opposed to each other with the crossing portion 26 interposed therebetween are extracted. Data on the position of the point sequence constituting the blood vessel walls 33, 34, 35, 36 is stored as coordinate values of a two-dimensional local coordinate system defined in the crossover region 28.

  In step S <b> 12, the vein analysis processing unit 10 of the image analysis apparatus 1 determines the position of the blood vessel wall of the vein 32 that has not been extracted because it intersects with the artery 30 to the position of the blood vessel wall 33, 34, 35, 36. Estimate using data and complement. Complementation of one blood vessel wall of the vein 32 at the intersection 26 is performed by connecting the position data of the point sequence constituting the blood vessel wall 33 and the position data of the point sequence constituting the blood vessel wall 35 by a spline. Done. Similarly, complementation of the other blood vessel wall of the vein 32 at the intersection 26 connects the position data of the point sequence constituting the blood vessel wall 34 and the data of the position of the point sequence constituting the blood vessel wall 36 by a spline. Is done by doing. In FIG. 5, the blood vessel wall 37 between the blood vessel wall 33 and the blood vessel wall 35 and the blood vessel wall 38 between the blood vessel wall 34 and the blood vessel wall 36 complemented by the vein analysis processing unit 10 are displayed. The figure is shown.

  Furthermore, the vein analysis processing unit 10 adds the estimated data of the positions of the blood vessel wall 37 and the blood vessel wall 38 to the data of the positions of the blood vessel walls 33, 34, 35, and 36, and the vein data is obtained from the data of these blood vessel walls. A distribution of 32 apertures is calculated (step S14). The distribution of the diameter of the vein 32 is calculated and output by the length of a perpendicular extending from one blood vessel wall 33, 35, 37 to the other blood vessel wall 34, 36, 38. An example of the distribution of the diameter of the vein 32 output by the vein analysis processing unit 10 is shown in FIG. Next, the vein analysis processing unit 10 calculates the average value of the diameters of veins belonging to portions other than the crossing part 26 and the minimum value of the diameters of the veins of the crossing part 26.

  Furthermore, the vein analysis processing unit 10 according to the present embodiment calculates the bending degree of the veins at the intersection 26 using the calculated data of the positions of the blood vessel wall 37 and the blood vessel wall 38. A method of calculating the degree of bending will be described with reference to FIG. The vein analysis processing unit 10 approximates the blood vessel wall 37 shown in FIG. 7A with two straight lines 45 and 46 as shown in FIG. 7B. The vein analysis processing unit 10 sets the angle θ1 formed by the straight lines 45 and 46 as the bending degree of the blood vessel wall 37. Similarly, the vein analysis processing unit 10 approximates the blood vessel wall 38 with a straight line 47 and a straight line 48, and sets an angle θ 2 formed by the two straight lines 47 and 48 as the bending degree of the blood vessel wall 38. Then, by calculating the average value of θ1 and θ2, the bending degree of the vein 32 at the intersection 26 is calculated (step S16).

  The vascular state determination processing unit 12 obtains the caliber of the venous caliber obtained as a result of examining the correlation between the result of analysis of various characteristics of the vein performed using a large number of fundus images and the presence or absence of suspicion of vascular abnormality. The reference value and the reference value of the bending degree are stored. These reference values indicate values that include approximately 80% of blood vessels that are not abnormal. In this embodiment, the reference value of the caliber of the vein stored in the blood vessel state determination processing unit 12 is the ratio of the minimum value of the caliber of the vein of the crossing portion 26 to the average caliber of the vein other than the crossing portion 26. Is a value expressed as a percentage. This reference value in the present embodiment is 60%. The reference value for the degree of bending is 120 °.

  The blood vessel state determination processing unit 12 uses the average diameter of the veins other than the intersection 26 calculated by the vein analysis processing unit 10 and the minimum value of the diameter of the intersection 26 to calculate the ratio of the minimum value to the average diameter. Ask. Then, the obtained ratio value is compared with a reference value of the caliber of the vein. In addition, the value of the degree of bending at the intersection 26 is compared with the reference value of the intersection. When the ratio of the minimum value of the diameter of the crossing portion to the average diameter is 60% or more and when the value of the bending degree is larger than 120 °, there is no abnormality in the vein 32 and the state of the vein is in a normal range. Judgment is made. If there is an abnormality in at least one of the ratio of the minimum value of the diameter of the crossing portion to the average diameter and the value of the degree of bending, it is determined that there is a suspicion of abnormality (step S18).

  When the processing of the blood vessel state processing unit 12 is finished, the display processing unit 16 performs the analysis of the fundus image 20, the distribution of the caliber and the degree of bending of the vein 32 calculated by the vein analysis processing unit 10, and the blood vessel state determination processing. The result of determination by the unit 12 is displayed on the output unit 18 (step S20). An example of the result display screen 40 displayed on the output unit 18 is shown in FIG. On the result display screen 40, the fundus image 20 that has been analyzed, the figure of the blood vessel wall extracted in the crossover region 28, the calculated distribution of the caliber of the vein, the degree of bending at the crossover, and the suspected abnormality The result of the judgment about the presence or absence is displayed. The ratio of the minimum value of the diameter of the crossing part 26 to the average diameter of the veins other than the crossing part 26 is displayed here as “minimum value of the diameter ratio of the crossing part”.

  The image analysis apparatus 1 according to the present embodiment can calculate the bending degree of the vein 32 and the distribution of the diameter of the vein 32 in the intersection 26 using the blood vessel wall data of the vein 32 extracted around the intersection 26. Then, by using the calculation results of the degree of bending and the distribution of the caliber, it is possible to make an objective and very reliable determination of whether or not there is a suspicion of a blood vessel abnormality based on quantitative characteristics.

  (Embodiment 2) FIG. 9 shows a block diagram of an image analyzing apparatus 41 in the present embodiment. The image analysis apparatus 41 according to the present embodiment is configured by a single computer, and includes a crossing area determination processing unit 42 and a vein wall extraction processing unit 44. Further, the image analysis device 41 includes an arteriovenous determination processing unit 6, a vein analysis processing unit 10, a blood vessel state determination processing unit 12, and a display processing unit 16 having the same configuration as the image analysis device 1 of the first embodiment. It has. These processing units included in the image analysis device 41 are stored in a format that can be executed in the storage unit of the computer, and the analysis of the fundus image proceeds by being executed by the CPU of the computer.

  In the computer constituting the image analysis device 41, the fundus image 20 displayed on the output unit 18 and the output unit 18 that can display the input fundus image and the result of the analysis performed by the image analysis device 41 are displayed. On the other hand, an input unit 14 including a pointing device such as a mouse that allows the examiner to specify the position of the crossover region 28 is provided. The image analysis apparatus 41 according to the present embodiment is characterized in that, in the first step of analysis, the examiner designates the crossing region 28 including the crossing portion 26 in the fundus image 20 to be analyzed.

  In the first step of image analysis by the image analysis apparatus 41 of the present embodiment, the examiner specifies the crossover region 28 including the crossover portion 26 to be analyzed while confirming the fundus image 20 displayed on the output unit 18. To do. The cross region determination processing unit 42 of the image analysis device 41 inputs coordinates on the fundus image 20 corresponding to the designated position, and sets the cross region 28 on the image 20. Here, when specifying the crossing region 28, the examiner specifies the crossing region 28 so that blood vessels other than the artery and vein crossing at the crossing portion 26 are not included, so that the following processing can be performed more quickly and accurately. Can proceed.

  The vein wall extraction processing unit 44 of the image analysis device 41 extracts the blood vessel wall of the vein 32 included in the crossing region 28. The vein wall extraction processing unit 44 first analyzes the pixel value of the crossover region 28 in detail to determine that the region having vein-specific chromaticity is a vein. Then, a pixel having a large amount of change in luminance value and having a luminance equal to or lower than a predetermined threshold is selected near the end of the region determined to be a vein, and this pixel is determined to be a vein wall. By extracting all such pixel groups and connecting these point groups, the vein wall is extracted. FIG. 10 shows an image of the crossing region 28 from which the vein blood vessel walls 33, 34, 35, and 36 are extracted.

  The vein analysis processing unit 10 of the image analysis apparatus 41 performs the same processing as in the first embodiment, and supplements with the spline using the data of the positions of the vein blood vessel walls 33, 34, 35, 36, and the crossing unit 26. The data of the position of the blood vessel walls 37 and 38 including are obtained. FIG. 11 shows a diagram of the crossing region 28 in which the vein blood vessel walls 37 and 38 are complemented. The vein analysis processing unit 10 calculates the distribution of the diameter of the vein 32 and the degree of bending at the intersection based on the data of the positions of the vein blood vessel walls 33 to 38 in the intersection region 28. Based on these calculation results, the blood vessel state determination processing unit 12 determines whether or not the vein 32 is suspected of being abnormal. The result of the determination is processed by the display processing unit 16 and displayed on the output unit 18 using the result display screen 40 as in the first embodiment.

  The image analysis apparatus 41 according to the present embodiment selects the crossover part 26 that the examiner most wants to determine whether or not the examiner is abnormal when the examiner designates the crossover region 28, and obtains the result. You can know quickly. Further, since the image analysis apparatus 41 of this embodiment extracts only the vein wall and does not extract the artery wall, the analysis result can be obtained more quickly.

  As mentioned above, although the specific example of this invention was described in detail in the Example, these are only illustrations and do not limit a claim. The technology described in the claims includes various modifications and changes of the specific examples illustrated above. For example, each processing unit of the image analysis apparatus in the present embodiment can be modularized and configured as an external device of the computer. The image analysis apparatus according to the present invention includes a fundus image photographing unit, so that image data can be input more quickly. Furthermore, the configuration of the screen output by the image analysis apparatus can be variously changed according to the use of the examiner.

1 is a block configuration diagram of an image analysis apparatus 1 according to Embodiment 1. FIG. It is a figure which shows the fundus image 20 analyzed in Examples 1 and 2. FIG. 6 is a flowchart illustrating the contents of blood vessel analysis processing executed by the image analysis apparatus 1 according to the first embodiment. It is a figure which shows the crossing area | region 28 by which the blood vessel wall was extracted. It is a figure which shows the crossing area | region 28 where the blood vessel wall of the vein 32 was complemented. It is a figure which shows the calculation result of distribution of the aperture of the vein. It is a figure which shows the state which approximated the blood vessel walls 37 and 38 in the crossing part 26 with the straight lines 45, 46, 47, and 48 for calculation of a bending degree. It is a figure of the result display screen 40 which displays the judgment result of the presence or absence of suspicion of abnormality. FIG. 6 is a block configuration diagram of an image analysis apparatus 41 according to a second embodiment. It is a figure which shows the crossing area | region 28 where the extraction of the vein wall was performed. It is a figure which shows the crossing area | region 28 where the blood vessel wall of the vein 32 was complemented.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1,41 Image analysis device 2 Blood vessel extraction processing unit 4,42 Crossing region determination processing unit 6 Arteriovenous determination processing unit 8 Blood vessel wall extraction processing unit
DESCRIPTION OF SYMBOLS 10 Vein analysis processing part 12 Blood vessel state judgment processing part 14 Input part 16 Display processing part 18 Output part 20 Fundus image 21 Optic nerve head 22 Retina 24 Blood vessel 26 Crossing part 28 Crossing area 30 Artery 32 Vein 33, 34, 35, 36, 37, 38 Blood vessel wall 40 Result display screen 44 Vein wall extraction processing unit 45, 46, 47, 48 Straight line approximating blood vessel wall

Claims (3)

An image analysis device for analyzing a crossing portion of a blood vessel captured in a fundus image obtained by photographing a fundus of a subject,
In the image, a crossing region determining means for determining a crossing region including a crossing portion of a blood vessel,
An arteriovenous determination unit that determines whether a blood vessel included in the crossover region is an artery or a vein; and a blood vessel wall extraction unit that extracts a blood vessel wall of a blood vessel included in the crossover region;
A vein analysis means for calculating the distribution and bending degree of the diameter of the vein in the crossover region from the extracted data of the position of the blood vessel wall of the vein,
The vein analysis means calculates the position of the blood vessel wall of the vein at the crossover portion based on the data of the position of the blood vessel wall extracted from the crossover region, and calculates the distribution of the caliber and the bending degree of the vein at the crossover portion. An image analysis apparatus characterized by the above. From the calculation result of the distribution of the caliber of the veins included in the crossover region and the degree of flexion, the blood vessel state determination means for determining whether there is a suspicion of a blood vessel abnormality in the fundus,
The image analysis apparatus according to claim 1, further comprising a display unit that displays a determination result of the blood vessel state determination unit. An image analysis program for analyzing the intersection of the blood vessels of the fundus from the data of the image obtained by photographing the fundus of the subject,
In the image, determine a crossover region including a crossing portion of the blood vessel,
Determine whether the blood vessel included in the crossover region is an artery or a vein;
Extracting the blood vessel wall of the blood vessel contained in the crossover region,
Calculate the extracted vein wall position data for veins,
From the calculated vascular wall position data, calculate the venous vascular wall position data at the intersection, and calculate the caliber distribution and bending degree of the venous diameter in the crossing area based on the calculated position data. An image analysis program for causing a computer to execute processing.