JP2008194334A - Endoscope image display method, device, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To simultaneously display an endoscope animation and a comparative endoscope animation, both showing the same internal body part. <P>SOLUTION: A blood vessel pattern X is extracted from the frame image PF of an endoscope animation P obtained by a pattern extraction means 3. In a shift vector operation means 5a, a shift vector showing a positional difference of blood vessels between a blood vessel pattern X and a comparative blood vessel pattern Y is generated, and the blood vessels within the blood vessel pattern X are shifted by the pattern shift means 5b using the shift vector. Next, a correlation coefficient ρ is calculated between the blood vessels X and all the comparative blood vessel pattern Y extracted from the comparative frame image PF0 constituting the comparative endoscope animations P0. In the comparative image selection means 5d, the comparative frame image PF0 including the blood vessel pattern X is selected, and both of the frame image PF including the blood vessel pattern X and the compared frame image PF0 are simultaneously displayed. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an endoscopic image display method, apparatus, and program for displaying an endoscopic moving image photographed and acquired using an endoscopic apparatus on a display device.

When taking an image of a body cavity of a subject using an endoscope device, information related to the endoscope image as well as an endoscope image taken using the endoscope device or a past image was taken. It has been proposed to display an endoscopic image simultaneously (for example, see Patent Document 1). In Patent Document 1, when an endoscope apparatus and an external connection device such as a personal computer are connected so as to be able to transfer data, and when it is desired to display a past endoscopic image of the same patient as the endoscopic image, the external connection is performed. It is possible to display a past endoscopic image retrieved by the device as a comparative image at the same time as the endoscopic image.
JP 2001-157200 A

  While the endoscopic image and the past endoscopic image displayed in Patent Document 1 are still images, in recent years, the captured endoscopic image may be displayed as a moving image. Therefore, it is desired to simultaneously display an endoscopic image and a past endoscopic image even when the endoscopic image is a moving image.

  However, as in Patent Document 1, when an endoscopic image and a past endoscopic image are displayed at the same time in the case of a moving image, an image of the same part in the body cavity is not always displayed on the display device at the same time. There is no problem. That is, even if the endoscopic image and the past endoscopic image are images of the same part in the body cavity, the same part is displayed between the endoscopic image and the past endoscopic image. May be different timing. On the other hand, in order to improve the diagnostic efficiency by comparing the endoscopic image and the past endoscopic image, the endoscopic image and the past endoscopic image when the same part is imaged are compared. It is necessary to display at the same time.

   Therefore, the present invention provides an endoscopic image display method, apparatus, and program capable of simultaneously displaying an endoscopic moving image and a comparative endoscopic moving image when a same part is imaged as a moving image on a display device. It is intended to provide.

  An endoscopic image display method according to the present invention extracts a blood vessel included in a frame image constituting an endoscope moving image of a subject acquired as a moving image using an endoscopic device as a blood vessel pattern, and stores the endoscope in advance. When selecting a comparison frame image including an extracted blood vessel pattern from a plurality of comparison frame images constituting a comparative endoscopic moving image of a subject acquired as a moving image using the apparatus and displaying the frame image on a display device The selected comparison frame image is displayed at the same time.

  An endoscope display apparatus according to the present invention includes a pattern extraction unit that extracts a blood vessel included in a frame image constituting an endoscope moving image of a subject acquired as a moving image using the endoscope apparatus as a blood vessel pattern, A pattern from an image database storing a comparative endoscopic moving image of a subject acquired as a moving image using an endoscopic device, and a plurality of comparison frame images constituting the comparative endoscopic moving image stored in the image database Comparison image processing means for selecting a comparison frame image including the blood vessel pattern extracted by the extraction means, and display control for simultaneously displaying the comparison frame image selected by the comparison image processing means when the frame image is displayed on the display device Means.

  The endoscopic image display program of the present invention extracts, as a blood vessel pattern, a blood vessel included in a frame image that constitutes an endoscopic moving image of a subject acquired as a moving image using an endoscopic device on a computer. A comparison frame image including an extracted blood vessel pattern is selected from a plurality of comparison frame images constituting a comparative endoscope moving image of a subject obtained as a moving image using an endoscope apparatus, and the frame image is displayed on a display device In this case, the selected comparison frame image is displayed at the same time.

  Here, the endoscope moving image may be an image of any part in the body cavity as long as it is a moving image acquired using the endoscope apparatus. Further, the comparative endoscopic moving image is not limited to the one obtained by photographing all the imaging parts taken by the endoscopic moving image, but may partially include the same subject as the endoscopic moving image. That's fine.

  Furthermore, the endoscope moving image may be displayed on the display device immediately after being acquired as a moving image using the endoscope device (real-time display), or once stored and stored in the image database. An endoscope moving image may be displayed on a display device.

  Further, the pattern extraction means may be any means as long as it extracts blood vessels as blood vessel patterns. For example, by performing binarization processing on the endoscopic image so that the blood vessels included in the endoscopic image become a black region, A blood vessel may be extracted as a blood vessel pattern, or feature quantities such as coordinates of a branch point, length, and branch angle of a vein pattern may be extracted as a blood vessel pattern.

  When the blood vessel pattern is composed of a binarized image, the comparison image processing unit may select the comparison frame image using the blood vessel pattern extracted from the frame image as it is, or the blood vessel pattern may be selected by the pattern rotation unit. A comparison frame image may be searched using a plurality of blood vessel patterns having different rotation angles generated by performing affine transformation on the image.

  The comparison image processing means may use any method as long as it selects a comparison frame image including a blood vessel pattern. For example, an optical flow estimation method may be used. A pattern recognition method may be used. When the comparison frame image is selected by the optical flow estimation method, for example, the comparison image processing unit causes the comparison image processing unit to detect a shift between the position of the black region in the blood vessel pattern and the position of the black region in the comparison blood vessel pattern. Shift vector calculation means for calculating as a shift vector, pattern shift means for shifting the position of the black area in the blood vessel pattern using the shift vector calculated by the shift vector calculation means, and the black area is shifted by the pattern shift means Coefficient calculation means for calculating the correlation degree between the blood vessel pattern and the comparison blood vessel pattern as a correlation coefficient, and a plurality of correlation coefficients calculated for each of the plurality of comparison blood vessel patterns included in each comparison frame image by the coefficient calculation means. Select the comparison frame image that contains the blood vessel pattern. May be such that a comparison image selection unit that.

  According to the endoscopic image display method, apparatus, and program of the present invention, a blood vessel included in a frame image constituting an endoscopic moving image of a subject acquired as a moving image using the endoscopic apparatus is extracted as a blood vessel pattern. Then, a comparison frame image including the extracted blood vessel pattern is selected from a plurality of comparison frame images constituting a comparative endoscope moving image of a subject acquired as a moving image in advance using an endoscope apparatus, and the frame image is displayed. When displaying the endoscopic moving image and the comparative endoscopic moving image simultaneously on the display device by simultaneously displaying the selected comparison frame image when displaying on the device, the endoscopic moving image and the comparative endoscopic image are displayed. Since the moving image displays the same part in the body cavity, diagnosis or the like can be performed while comparing the endoscopic moving image and the comparative endoscopic moving image, thereby improving the diagnostic efficiency. Can.

  The image processing apparatus further includes pattern rotation means for generating a plurality of blood vessel patterns having different rotation angles by performing affine transformation on the blood vessel pattern, and the comparison image processing means uses the blood vessel pattern generated by the pattern rotation means. When selecting a frame image including the blood vessel pattern, when obtaining an endoscopic moving image and a comparative endoscopic moving image by photographing, the blood vessel pattern is obtained even when the images are taken from different photographing angles. The comparison frame image including it can be searched.

  Further, the comparison image processing means is calculated by the shift vector calculation means and the shift vector calculation means for calculating the shift between the position of the black area in the blood vessel pattern and the position of the black area in the comparison blood vessel pattern as a shift vector. Coefficient calculation that calculates the correlation coefficient between the pattern shift means that shifts the position of the black area in the blood vessel pattern using the shift vector and the blood vessel pattern whose black area is shifted by the pattern shift means and the comparative blood vessel pattern as a correlation coefficient And a comparison image selection means for selecting a comparison frame image including a blood vessel pattern using a plurality of correlation coefficients respectively calculated for a plurality of comparison blood vessel patterns included for each comparison frame image in the coefficient calculation means. Search for comparison frame images with high accuracy and speed. Door can be.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. Hereinafter, preferred embodiments of an endoscope display device of the present invention will be described in detail with reference to the drawings. The configuration of the endoscopic image display device 1 as shown in FIG. 1 is realized by executing an endoscopic image display program read into the auxiliary storage device on a computer (for example, a personal computer). At this time, this endoscopic image display program is stored in an information storage medium such as a CD-ROM, or distributed via a network such as the Internet and installed in a computer.

  The endoscopic image display device 1 simultaneously displays an endoscopic moving image P of a subject acquired by the endoscopic device 50 and an endoscopic moving image P obtained when the inside of the subject's body cavity is captured in the past. When displaying on the display device 100, the endoscopic moving image P and the comparative endoscopic moving image P0 are displayed on the display device 100 in the same body cavity. The endoscopic image display device 1 includes an image acquisition unit 2, a pattern extraction unit 3, an image database DB, a comparison image processing unit 5, a display control unit 6, and the like.

  The image acquisition unit 2 acquires the endoscope moving image P in the body cavity of the subject acquired by the endoscope apparatus 50. Note that the image acquisition unit 2 may acquire the video imaged in the endoscope apparatus 50 in real time, or has already been captured by the endoscope apparatus 50 and is stored in the image database DB. The endoscope moving image P may be acquired.

  The pattern extraction unit 3 extracts the blood vessel pattern X included in each frame image PF of the endoscope moving image P acquired by the image acquisition unit 2. Specifically, the pattern extraction means 3 performs a filtering process for binarizing (monochrome) the color information of the frame image PF. Then, the blood veins in the frame image PF appear as black areas in the binarized image. The pattern extraction unit 3 extracts a binarized image in which a blood vessel appears as a black region as a blood vessel pattern X.

  The image database DB stores a comparative endoscope moving image P0 composed of a plurality of comparison frame images PF0 obtained by photographing a subject using an endoscope apparatus in advance and acquired as a moving image. This comparative endoscopic moving image P0 is obtained by photographing the same subject as the above-described endoscopic moving image P, and is obtained, for example, by photographing the subject's body cavity (for example, the esophagus) in the past. It is made up of. In the image database DB, the comparative blood vessel pattern Y extracted by the pattern extracting means 3 described above is stored in association with each comparative frame image PF0. The image database DB stores a comparative endoscope moving image P0 and a comparative blood vessel pattern Y of a plurality of subjects. For example, the pattern selection means DBM performs authentication database DB according to input of personal information such as a keyboard and a mouse. The comparative endoscopic moving image P0 associated with the personal information PD input from is selected, and the comparative blood vessel pattern Y of the comparative endoscopic moving image P0 is selected.

  The pattern rotation means 4 generates a plurality of blood vessel patterns X having different rotation angles by converting the blood vessel pattern X into, for example, 15 ° by affine transformation. As described above, by rotating the blood vessel pattern X, even when the imaging angle (field angle) and the imaging direction of the blood vessel pattern X and the comparative blood vessel pattern Y are different, the verification can be performed reliably.

  The comparison image processing means 5 collates the blood vessel pattern X extracted by the pattern extraction means 3 with the comparison blood vessel pattern Y of a specific person using the Lucas-Kanade method which is a kind of optical flow estimation method, and It is verified that the person who is the subject of the moving image P is a specific person.

  Specifically, as shown in FIG. 2, the comparison image processing unit 5 includes a shift vector calculation unit 5a, a pattern shift unit 5b, a coefficient calculation unit 5c, a comparison image selection unit 5d, and the like. The Lucas-Kanade method was devised for recognizing the same point of an object in images taken simultaneously for stereoscopic viewing with two cameras. This is based on the assumption that the object in the image is hardly deformed. The shift vector calculation method is disclosed in 'Pyramidal Implementation of the Lucas Kanade Feature Tracker Description of the algorithm'.

  The shift vector calculation means 5a calculates a shift between the position of the blood vessel (black region) in the blood vessel pattern X and the position of the blood vessel (black region) in the comparative blood vessel pattern Y as a shift vector. That is, when obtaining the parallel movement distance of the blood vessel pattern X in the binarized image, it is assumed that there is little deformation of the object between the extracted blood vessel pattern X and the comparative blood vessel pattern Y, and the Lucas-Kanade method is used. A shift vector is calculated. In other words, the shift vector between the two blood vessel patterns X and Y is calculated under the assumption that the same luminance pixels in the minute region are shifted (moved) in the same blood vessel pattern in the two images. To do. Then, the pattern shift unit 5b shifts the position of the blood vessel (black region) in the blood vessel pattern X using the shift vector calculated by the shift vector calculation unit 5a. If the blood vessel pattern X and the comparative blood vessel pattern Y are not similar, an inaccurate shift amount is calculated. However, since the degree of coincidence of the correlation coefficient ρ described later is low, it is removed from the coincidence candidates at that time. Will do.

  Further, the shift vector is calculated by the shift vector calculation means 5a after the blood vessel pattern X is rotated, but the shift vector is too large after calculating the shift vector from the blood vessel pattern X and the comparative blood vessel pattern Y. Alternatively, the blood vessel pattern X may be rotated by the pattern rotating means 4.

The coefficient calculating means 5c calculates the correlation degree between the blood vessel pattern X whose black region has been shifted by the pattern shifting means 5b and the plurality of comparative blood vessel patterns Y stored in the image database DB as a correlation coefficient ρ. Specifically, the coefficient calculating means 5c divides the covariance of each blood vessel pattern X and Y by the product of the standard deviation calculated for each blood vessel pattern X and Y as shown in the following formula (1), for example. Thus, the correlation coefficient ρ is calculated.

  Here, Xi (i = 1, 2,... N, N is the number of pixels of the blood vessel pattern X) is the pixel value of the blood vessel pattern X, and Yi (i = 1, 2,... N, N is the blood vessel pattern. Y represents the pixel value of the comparative blood vessel pattern Y. Note that the blood vessel pattern X and the comparative blood vessel pattern Y are each formed by arranging pixels in a two-dimensional direction. However, in the above equation (1), the correlation coefficient ρ is rearranged into N one-dimensional pixel rows. Is calculated. Specifically, the coefficient calculating unit 5c connects the pixel rows from the upper left to the upper right of the blood vessel patterns X and Y from the upper row to the lower row, and rearranges the data into data composed of N one-dimensional pixel rows. Then, the correlation coefficient ρ is calculated by the equation (1). By calculating the correlation coefficient ρ as shown in the above equation (1), it is possible to compare the two blood vessel patterns X and Y without being affected by the overall average brightness change.

  The comparison image selection unit 5d selects the comparison frame image PF0 including the blood vessel pattern X by using the plurality of correlation coefficients ρ respectively calculated for the plurality of comparison blood vessel patterns Y by the coefficient calculation unit 5c. Here, the comparison image selection unit 5d has little correlation when the correlation coefficient ρ = 0.0 to 0.2, for example, and has a slight correlation when the correlation coefficient ρ = 0.2 to 0.4. When the correlation coefficient ρ = 0.4 to 0.7, there is a considerable correlation, and when the correlation coefficient ρ = 0.7 to 1.0, it is determined that there is a strong correlation. If there is, for example, a correlation coefficient ρ = 0.7 to 1.0 among the plurality of correlation coefficients ρ calculated for one blood vessel pattern X, the comparison image selection means 5d has the highest correlation degree. A strong comparison frame image PF0 is selected.

  On the other hand, for example, when all of the plurality of correlation coefficients ρ are correlation coefficients ρ <0.7, the comparison image selection unit 5d does not select a frame image, or the correlation coefficient ρ <0.7. Even in this case, the comparison frame image PF0 having the largest correlation coefficient is selected.

  The display control means 6 in FIG. 1 simultaneously displays the comparison frame image PF0 selected by the comparison image processing means 5 when displaying the frame image PF on the display device 100. As a result, the display device 100 simultaneously displays the endoscope moving image P and the comparative endoscope moving image P0 acquired by photographing substantially the same part in the body cavity. That is, every time the frame image PF displayed on the display device 100 is sequentially switched when the endoscope moving image P is displayed as a moving image, the comparison frame image PF0 is also switched accordingly. As a result, even in the case of a moving image, the endoscope device 50 is used to perform an endoscopic observation of a predetermined part in the body cavity while referring to the comparative endoscope moving image P0 obtained when a predetermined part is photographed in the past. Since the mirrored image P can be diagnosed, the diagnostic efficiency can be improved.

  Although the display control means 6 displays the frame image PF and the comparison frame image PF0 itself as they are, they may be displayed after processing the frame image PF. For example, the frame image PF or the comparison frame image PF0 is processed and displayed on the display device 100 using a known technique such as marking and displaying a difference between the frame image PF and the selected comparison frame image PF0. Anyway.

  Moreover, the display control means 6 may display the endoscope moving image P photographed and acquired by the endoscope apparatus 50 in real time, or an endoscope moving image stored in the image database DB. The image P may be displayed. At this time, the frame image PF and the comparison frame image PF0 selected by the comparison image selection means 5d are stored in the image database DB in association with each other.

  FIG. 3 is a flowchart showing a preferred embodiment of the endoscopic image display method of the present invention, and the endoscopic image display method will be described with reference to FIGS. 1 to 3. First, imaging within the body cavity is performed by the endoscope apparatus 50, and the endoscope moving image P is acquired by the image acquisition means 2 (step ST1). Next, the frame image PF of the endoscope moving image P acquired by the pattern extraction unit 3 is binarized, whereby a blood vessel composed of a binarized image in which blood vessels in the frame image PF are expressed as black regions. Pattern X is extracted (step ST2).

  Then, the pattern rotation means 4 performs affine transformation on the extracted blood vessel pattern X, thereby generating a plurality of blood vessel patterns X having different rotation angles (step ST3).

  On the other hand, when personal information is input from a keyboard, a mouse, or the like, an endoscope moving image P0 of a subject to be authenticated is selected from the image database DB by the image selection means DBM, and each comparison frame of the endoscope moving image P0 is selected. The comparison blood vessel pattern Y associated with the image PF0 is sequentially selected (step ST4). The comparative blood vessel pattern Y may be acquired before the generation of the shift vector. For example, it may be performed before the acquisition of the endoscope image P (step ST1).

Thereafter, in the shift vector calculation means 5a, a shift vector indicating a blood vessel position shift between the rotated blood vessel pattern X and the comparative blood vessel pattern Y is generated, and the black region in the blood vessel pattern X uses the shift vector by the pattern shift means 5b. (Step ST5).

  Next, in the coefficient calculation means 5c, the correlation coefficient ρ between the shifted blood vessel pattern X and the comparative blood vessel pattern Y is calculated using the above equation (1) (step ST6). The calculation of the correlation coefficient ρ is calculated between the plurality of rotated blood vessel patterns X and all of the comparative blood vessel patterns Y extracted from the comparative frame image PF0 constituting the comparative endoscope moving image P0. (Steps ST4-7).

  Thereafter, in the comparison image selection means 5d, it is determined whether or not the correlation coefficient ρ is 0.7 or more among the plurality of correlation coefficients ρ calculated for one blood vessel pattern X (step ST8). When there is a comparison frame image PF0 of 0.7 or more among the plurality of correlation coefficients ρ, it is determined that the comparison frame image PF0 includes the blood vessel pattern X, and this comparison frame image PF0 is displayed on the display device 100. Selected as a thing. Then, the display control means 6 simultaneously displays the frame image PF including the blood vessel pattern X and the comparison frame image PF0 (step ST9). On the other hand, when all of the plurality of correlation coefficients ρ are less than 0.7, it is determined that there is no comparison frame image PF0 including the blood vessel pattern X, and the display control means 6 displays the comparison frame image PF0 on the display device 100. Not performed (step ST10).

  According to the above embodiment, when displaying an endoscope moving image P composed of a plurality of frame images PF obtained by capturing an image of a subject in a body cavity using the endoscope device 50 as a moving image on the display device, A blood vessel included in the frame image PF of the mirror image P is extracted as a blood vessel pattern X, and is extracted from a plurality of comparison frame images PF0 constituting the comparative endoscope moving image P0 obtained by photographing a subject in advance and acquired as a moving image. By selecting the comparison frame image PF0 including the blood vessel pattern X and simultaneously displaying the comparison frame image PF0 selected when the frame image PF is displayed on the display device 100, the frame image PF displayed on the display device 100 is captured. Since the comparison frame image PF0 obtained by photographing the same part as the part can be displayed at the same time, the endoscope moving image P and the comparison endoscope moving image P0 While comparing can be improved diagnostic efficiency by performing the diagnosis.

  The blood vessel pattern X is further provided with pattern rotation means 4 for generating a plurality of blood vessel patterns X having different rotation angles by performing affine transformation, and the comparison image processing means 5 is a blood vessel generated by the pattern rotation means 4. When searching for a frame image PF including a blood vessel pattern X using a pattern, when acquiring an endoscopic moving image P and a comparative endoscopic moving image P0 by shooting, when shooting from different shooting angles, etc. Even if it exists, the comparison frame image PF0 including the blood vessel pattern X can be searched.

  In addition, the comparison image processing unit 5 calculates, as a shift vector, a shift between the position of the blood vessel pattern in the frame image PF and the position of the comparative blood vessel pattern Y extracted from the comparison frame image PF, A pattern shift means 5b for shifting the position of the blood vessel pattern X in the frame image PF using the shift vector calculated by the shift vector calculation means 5a, a frame image PF in which the blood vessel pattern X is shifted by the pattern shift means 5b, and a comparison frame A coefficient calculation unit 5c that calculates the degree of correlation with the image PF0 as a correlation coefficient ρ, and a comparison including the blood vessel pattern X using the correlation coefficient ρ calculated for each of the plurality of comparative blood vessel patterns Y by the coefficient calculation unit 5c. Comparison image selection means 5d for selecting the frame image PF0; The comparison frame image PF0 can be searched with high accuracy and high speed.

  The embodiment of the present invention is not limited to the above embodiment. For example, the comparison image processing unit 5 selects and associates the comparison frame image PF0 using the blood vessel pattern X and the comparison blood vessel pattern Y rotated by the pattern rotation unit 4, but not the blood vessel pattern X. The correlation coefficient ρ with the comparative blood vessel pattern Y may be calculated using the extracted blood vessel pattern X itself, and the comparison frame image PF0 may be selected and associated.

  In addition, the blood vessel pattern X is extracted as an image in the pattern extraction unit 3 of FIG. 1, but feature quantities such as the coordinates, length, and branch angle of the branch point of the vein pattern are extracted as the blood vessel pattern X. Anyway. At this time, feature quantities such as the coordinates of the branch points of the vein pattern described above are also stored in the image database DB, and the comparison image processing means 5 can perform matching using known pattern recognition using these feature quantities. Will be done.

  In the above embodiment, the pattern rotation means 4 rotates the extracted blood vessel pattern X. However, the comparative blood vessel pattern Y stored in the image database DB may be rotated by affine transformation. . Alternatively, the rotated comparison blood vessel pattern Y may be stored in the image database DB in advance, and the rotation processing by the pattern rotation unit 4 may not be performed when the comparison image processing unit 5 performs collation.

  Furthermore, in the above-described embodiment, the case where it is performed for all the frame images PF of the endoscope moving image P is illustrated, but the above-described comparison frame image PF0 is not selected for all the frame images PF. When the endoscope moving image P is being reproduced, the frame image PF is sampled at regular time intervals, and the above-described comparison frame image PF0 is searched for the sampled frame image PF. The comparative endoscope moving image P0 may be reproduced.

  Further, although a case where a moving image is displayed is illustrated, the above-described matching process is performed on the endoscope moving image P and the comparative endoscope moving image P0 in advance, and the frame image PF and the comparative frame image PF0 are displayed. Are stored in association with each other, the endoscope moving image P and the comparative endoscope moving image P0 are stored, and when the endoscope moving image P is reproduced, the comparative endoscope moving image P0 is synchronized with this. May be reproduced and displayed on the display device 100 at the same time.

The block diagram which shows preferable embodiment of the endoscopic image display apparatus of this invention The block diagram which shows an example of the collation means in the endoscopic image display apparatus of FIG. The flowchart which shows preferable embodiment of the endoscopic image display method of this invention

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Endoscopic image display apparatus 2 Image acquisition means 3 Pattern extraction means 4 Pattern rotation means 5 Comparison image processing means 5a Shift vector calculation means 5b Pattern shift means 5c Coefficient calculation means 5d Comparison image selection means 6 Display control means 50 Endoscope Device 100 Display device P Endoscopic moving image P0 Comparative endoscopic moving image PF Frame image PF0 Comparative frame image X Blood vessel pattern Y Comparative blood vessel pattern ρ Correlation coefficient

Claims (6)

Extracting a blood vessel pattern included in a frame image constituting an endoscope moving image acquired as a moving image from a subject using an endoscope apparatus;
Select the comparison frame image including the extracted blood vessel pattern from a plurality of comparison frame images constituting a comparative endoscope moving image acquired as a moving image from the subject using the endoscope device in advance,
An endoscopic image display method, comprising: simultaneously displaying the selected comparison frame image when displaying the frame image on the display device. Pattern extraction means for extracting a blood vessel included in a frame image constituting an endoscope moving image of a subject acquired as a moving image using an endoscope apparatus as a blood vessel pattern;
An image database storing the comparison endoscope moving image of the subject acquired as a moving image in advance using the endoscope device;
Comparison image processing means for selecting the comparison frame image including the blood vessel pattern extracted by the pattern extraction means from the plurality of comparison frame images constituting the comparison endoscope moving image stored in the image database; ,
An endoscopic image display device comprising: display control means for simultaneously displaying the comparison frame image selected by the comparison image processing means when the frame image is displayed on the display device.   The pattern extraction means extracts the blood vessel as the blood vessel pattern by performing binarization processing on the endoscopic image so that the blood vessel included in the endoscopic image becomes a black region. The personal authentication device according to claim 2.   The apparatus further includes pattern rotation means for generating a plurality of blood vessel patterns having different rotation angles by performing affine transformation on the blood vessel pattern, and the comparison image processing means is configured to output the blood vessel pattern generated by the pattern rotation means. 4. The endoscopic image display device according to claim 3, wherein the frame image including the blood vessel pattern is used to select the frame image. The comparative image processing means is
Shift vector calculating means for calculating a shift between the position of the black region in the blood vessel pattern and the position of the black region in the comparative blood vessel pattern as a shift vector;
Pattern shifting means for shifting the position of the black region in the blood vessel pattern using the shift vector calculated by the shift vector calculating means;
Coefficient calculating means for calculating a correlation coefficient between the blood vessel pattern in which the black region is shifted by the pattern shifting means and the comparative blood vessel pattern as a correlation coefficient;
Comparison image selection means for selecting the comparison frame image including the blood vessel pattern by using the plurality of correlation coefficients respectively calculated for the plurality of comparison blood vessel patterns included in each of the comparison frame images in the coefficient calculation means. The endoscopic image display device according to claim 3 or 4, characterized by comprising: On the computer,
Extracting a blood vessel included in a frame image constituting an endoscope moving image of a subject obtained as a moving image using an endoscope device as a blood vessel pattern;
Selecting the comparison frame image including the extracted blood vessel pattern from a plurality of comparison frame images constituting a comparison endoscope moving image of the subject acquired as a moving image in advance using the endoscope apparatus;
An endoscope image display program for executing the simultaneous display of the selected comparison frame image when displaying the frame image on the display device.

Images