JPH0363029A - Image processor for ophthalmic use - Google Patents

Abstract

PURPOSE:To enable a measurement to be carried out objectively by detecting a concentration distribution between specified two points in a fundus image having multistage concentration levels, by converting the fundus image into a binary-coded image through setting a threshold value and by finding out a width of an identical binary-coded concentration range on a line linking the two points. CONSTITUTION:A fundus retinal image is picked up with a CCD camera 2 arranged in a fundus camera 1, the image is digitized by a converter 3 and stored in an image memory 6. The stored fundus image is displayed on a monitor 8 and two end points A, B of a vascular image including reflection region on a blood vessel 18 are specified by moving a cursor with a mouse 7. Next, a concentration distribution curve between the two points A and B is displayed on the monitor 8, a line (l) passing horizontally on the curve is set up, a binary-coded image is formed by making a concentration corresponding to height of the line (l) set up as a threshold value and the binary-coded image is displayed. Then the threshold value is varied in accordance with up and down of the mouse 7 and the inary-coded image is also varied. After a condition, wherein the binary-coded image indicates a total image of the blood vessel alone, is attained, numbers of picture elements of identical concentrations are counted and distances corresponding to the numbers of picture elements are calculated.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an ophthalmological image processing device that measures the width of an arterial reflection zone on the retina of the fundus in, for example, a digital fundus image.

[Prior Art] Conventionally, when measuring the width of the reflection area at the center of a blood vessel or the size of a lesion seen in retinal arteriosclerosis, the measurer has to view the fundus image on a slide viewer. However, the range of the area to be measured was determined, and measurements were taken using calipers, etc.

[Problems to be Solved by the Invention] However, in fundus images, the boundaries between reflective areas of blood vessels or lesions and other areas are generally not clear. However, it was difficult to always perform the same measurements, and the measurement results lacked objectivity.

[Means for Solving the Problems] The present invention that solves the above-mentioned problems includes a means for specifying two desired points in a fundus image having multiple density levels, and a means for specifying the density distribution between the two specified points. means for detecting, means for setting a threshold based on maximum and minimum values of density within the density distribution, means for converting the fundus image into a binarized image using the threshold, and the binarized image. The ophthalmological image processing apparatus is characterized in that it has means for determining the width of the binarized same density area on the line connecting the two points.

[Embodiment] FIG. 1 is a block diagram of an embodiment of the present invention, in which 1 is a fundus camera, 2 is a CCD camera, 3 is an A/D converter, 4 is a slide scanner, 5 is an image processing processor, 6 is an image memory, 7 is a mouse that sends instructions to the image processing processor 5, and 8 is a monitor that displays images, text information, etc. In addition to the mouse, various input means such as a keyboard, digitizer, trackball, etc. can be used as the instruction means.

FIG. 2 is a flowchart showing the processing procedure of this embodiment. First, the CCD camera 2 attached to the fundus camera 1 captures a fundus retinal image of the eye to be examined, and the obtained fundus image is digitized via the A/D converter 3 and stored in the image memory 6 (Fig. 2- 9). Although the fundus image is a color image in this embodiment, it may be a black and white image. The natural fundus image can be input by inputting a 35 mm color slide using the slide scanner 4, or by A/D converting the output of an analog medium such as a still video or a video recorder.

Next, display the fundus image stored in the image memory 6 on the monitor 8 as shown in FIG. 3, move the cursor on the monitor with the mouse 7, and as shown in FIG. Specify the two end points (blood vessel wall) A and B of the blood vessel image that includes the reflection area you want to measure above (Figure 2-1
0). Note that points A and B are not limited to the blood vessel wall, and may be any two points on a straight line passing through them.

Next, a curve representing the concentration distribution between the two points A and B is displayed on the monitor 8 as shown in the lower part of Fig. 4, and a line ℃ passing horizontally on this concentration distribution curve is set. The density corresponding to the height of the line °C is set as a threshold, and a binarized image is formed on the upper and lower sides divided into two by this threshold.
A binary image is also displayed on top (Figure 2-11).

Next, as the measurer moves the mouse 7 up and down, the line A is moved up and down to change the threshold value.

At this time, when the line °C is moved, the binarization threshold changes accordingly, and the binarized image also changes. This is processed and displayed in real time. The real-time binarization process and the pixel count measurement process of the blood vessel diameter 22 that are necessary here are executed by the image processing processor 5.

Here, when the position of ℃1 in FIG. 5(a), that is, the binarized image in FIG. In response to this instruction, the image processing processor counts the number of pixels of the same density on the line connecting the two specified points, that is, the number of pixels representing the blood vessel diameter 22, and calculates the distance according to the number of pixels. Figure 2-12, 13).

Similarly, regarding the width of the reflective part in the blood vessel image, the operator moves the line C in Fig. 4 up and down using the mouse, and
The position of I12 in Figure (b), that is, the binary image in Figure 4 is the third
When only the reflective part of the blood vessel image 18 shown in the figure is displayed, click the button of the mouse 7 again to count the number of pixels with the same density between two specified points corresponding to the width 23 of the reflective part of the blood vessel image. , calculate the distance according to the number of pixels (Fig. 2-14, 15).

The blood vessel diameter and the reflection width measured by the above processing have a relationship as shown at 22.23 in FIG. 6, respectively. That is, points C and D in FIG. 5(a) are respectively A and D in FIG.
Since it corresponds to B, ℃.

A line segment 22 cut along the line represents the blood vessel diameter 22. Further, the reflection width 23 is the darkest part in FIG. 5(b), that is, points E and G corresponding to the minimum value of density, and the brightest part (center of reflection), that is, points F corresponding to the maximum value of density. Since it corresponds to the distance 23 connecting two points located in the middle of the points, the inner line segment 23 cut by the line of 0.degree. C. 2 represents the width of the reflective section.

Note that the processing after the measurer specifies the desired points A and B can be automated without the measurer's instructions. In this case, calculate the maximum value F, minimum value E, and G of the concentration distribution curve between points C and D in Figure 5, and then calculate the intermediate value between E and F, and the intermediate value between F and G. automatically calculate f12. The intermediate value may be a density that is half the sum of the local maximum value and the local minimum value, or may be an intermediate value of a predetermined ratio other than that. Alternatively, it may be the same value as the maximum value or the minimum value.

After determining the blood vessel diameter and the width of the reflective part as described above,
To calculate the ratio of the blood vessel diameter 22 and the reflection width 23, actually r
atio=reflection width/vessel diameter), and the results are displayed on the monitor 8 (FIG. 2-16, 17). At this time, if the scale has been set in advance, it is also possible to output the results as absolute values.

In addition, as another method for determining the blood vessel diameter 22 and the reflection width 23, in FIG.
Specify point D, extract blood vessel diameter 22, and then extract E, F,
The reflection width 23 can also be extracted by specifying point G and connecting the midpoint between E and F and the midpoint between F and G.

As explained above, in a digital fundus image with multiple density levels, the shape of the density distribution curve is Since the width of the reflective part can be easily extracted from the measured value, and since the measurement is performed using a fixed algorithm, stable measurement values can be obtained without the subjectivity of the measurer.

In addition, the diameter of the blood vessel can be measured using the same algorithm as the width of the reflection area, so by finding the ratio between the width of the reflection area and the diameter of the blood vessel, it is possible to measure the width of the reflection area without being affected by magnification. As a result, comparisons with other reflection sites can be easily made regardless of the difference in magnification.

Note that the present invention is applicable not only to the measurement of blood vessels as described above but also to, for example, the measurement of the size of a lesion on the fundus of the eye.

[Effects of the Invention] According to the present invention, the width of a desired region on a fundus image can be visually determined.

[Brief explanation of drawings]

Figure 1 is the constituent countries of the embodiment of the present invention, Figure 2 is a flowchart showing the processing procedure of the embodiment, Figure 3 is a human fundus image including an arterial reflection image, and Figure 4 is the binarization of the human fundus image. The processed images, FIGS. 5(a) and 5(b), are explanatory diagrams of the procedure for processing the density distribution between two points, and FIG. 6 is a diagram showing the relationship between the blood vessel diameter and the width of the reflective part,
The main symbols in the diagram are: 1. Fundus camera, 2. CCD camera, 3. A/D converter, 4. Slide scanner, 5. Image processing processor. , 6...image memory, 7...mouse, 8...monitor, diagram (λ) (b) 2.

Claims (3)

[Claims] (1) Desired 2 images in a fundus image with multiple density levels
means for specifying a point; means for detecting a density distribution between the two specified points; means for setting a threshold based on the maximum and minimum values of density within the density distribution; An ophthalmological image processing apparatus comprising: means for converting into a digitized image; and means for determining the width of the binarized same density area on a line connecting the two points in the binarized image. (2) The specified two points are end points representing the outer diameter of the blood vessel image in the fundus image, and the reflection part and other points in the blood vessel image are converted into a binarized image, and the width of the reflection part of the blood vessel is determined. Section (1)
The ophthalmological image processing device described. (3) Claim (2) in which the ratio between the outer diameter of the blood vessel determined from the distance between the two designated points and the width of the reflective part of the blood vessel is determined.
The ophthalmological image processing device described.