JP4668624B2 - Image processing device for esophageal mucosa - Google Patents

Description

  The present invention relates to an image processing apparatus for esophageal mucosa that performs determination of a Barrett's esophagus or the like caused by gastroesophageal reflux by analyzing a characteristic amount of a mucosal boundary in the esophagus, more specifically, an epithelial boundary.

In recent years, endoscope apparatuses that perform endoscopy and the like using an endoscope have been widely adopted in the medical field and the industrial field. In the medical field, for example, by inserting an insertion portion of an endoscope into a body cavity and observing the examination target part, it is diagnosed whether the examination target part is in a normal state or a property change state in which the property has changed. Used for
In such a case, when it is possible to determine whether a normal state or a property change state by image processing from an endoscopic image, the surgeon focuses on the determination result site. Efficient diagnosis can be performed.
For example, Japanese Patent Application Laid-Open No. 2000-155840 as a conventional example describes that feature amounts are calculated based on information on the shape of the boundary between a property change portion and a normal portion while reducing the influence of imaging conditions.

  PCT International Publication No. WO02 / 073507 A2 discloses a method and an apparatus for performing abnormality detection using colorimetry or the like in image information obtained by a capsule endoscope.

Further, in screening for an esophageal examination as a tubular organ by an endoscope apparatus, the presence or absence of Barrett mucosa or Barrett esophagus is examined. Barrett's mucosa is a gastric esophageal junction at the junction of the esophagus and stomach where squamous epithelium forming the esophageal mucosa is replaced by gastric mucosa due to the influence of reflux esophagitis or the like, and is also called columnar epithelium. A disease called Barrett's esophagus is diagnosed when the Barrett's mucosa occurs 3 cm or more from the normal mucosal boundary and circumferentially with respect to the esophageal lumen cross section.
Barrett's esophagus is increasing, especially in Westerners, and is a major problem because adenocarcinoma occurs with high probability, so early detection of Barrett's esophagus or Barrett's mucosa is very important.
The Barrett's esophagus or Barrett's mucosa may not only occur all around the esophagus, but often progress locally.
JP 2000-155840 A PCT International Publication WO02 / 073507 A2

In the conventional examples according to the above two publications, no suggestion is made regarding the detection of the epithelial boundary to determine the Barrett esophagus or Barrett mucosa.
In addition, in the case of an image obtained by imaging a tubular part such as the inside of the esophagus with a direct-viewing endoscope, an image obtained by imaging the tubular mucosal tissue (epithelium) from the axial direction of the lumen is obtained. The shape and the like at the portions where the distances are different greatly change.

For this reason, it may take time to grasp the shape, size, and the like only with an endoscopic image captured by a direct-viewing type endoscope as in the conventional example. For this reason, it is very convenient to provide a determination result and provide the determination result together with an image that is easier to grasp.
(Object of invention)
The present invention has been made in view of the above-described points, and can efficiently diagnose Barrett's esophagus even in the case of an endoscopic image of the esophagus imaged by a direct-viewing endoscope. An object is to provide a suitable image processing apparatus for esophageal mucosa.

The first image processing apparatus for esophageal mucosa according to the present invention includes a development view generating means for generating a development view for an image taken in the esophagus, a color tone change of a pixel value on the development view, or an edge of the image based on the state, configuration and squamous as esophageal side mucosa epithelial boundary detecting means for detecting epithelial boundary as the boundary between the columnar epithelium of gastric side mucosa, the epithelium boundary detected by the epithelial boundary detecting means And a feature amount calculating unit that calculates a feature amount related to position information of each pixel, and an analysis unit that outputs an analysis result based on the feature amount calculated by the feature amount calculating unit .
The second image processing apparatus for esophageal mucosa according to the present invention is a development drawing generation means for generating a development drawing for an image captured in the esophagus, a color tone change of a pixel value on the development drawing, or an edge of the image An epithelial boundary detecting means for detecting an epithelial boundary as a boundary between the squamous epithelium as the esophageal mucosa and the columnar epithelium of the gastric mucosa based on the state, and the epithelial boundary detected by the epithelial boundary detecting means A feature amount calculating unit that calculates a feature amount related to coordinates of each pixel in the lumen direction; and an analysis unit that outputs an analysis result based on the feature amount calculated by the feature amount calculating unit. To do.
The third image processing apparatus for esophageal mucosa according to the present invention is a development diagram generating means for generating a development diagram for an image captured in the esophagus, a color change of a pixel value on the development diagram, or an edge of the image An epithelial boundary detecting means for detecting an epithelial boundary as a boundary between the squamous epithelium as the esophageal mucosa and the columnar epithelium of the gastric mucosa, based on the state, and a reference part setting means for setting the reference part on the developed view And a feature amount calculation for calculating a feature amount relating to a distance between the pixels constituting the epithelial boundary detected by the epithelial boundary detection means and the coordinates of the pixels constituting the reference portion set by the reference portion setting means And an analysis means for outputting an analysis result based on the feature quantity calculated by the feature quantity calculation means.

  According to the present invention, an analysis result for Barrett's esophagus and the like can be obtained, and diagnosis and the like can be advanced efficiently from the analysis result and the generated developed view.

  Embodiments of the present invention will be described below with reference to the drawings.

1 to 12 relate to a first embodiment of the present invention, FIG. 1 shows a configuration of an endoscope system including the first embodiment of the present invention, and FIG. 2 shows a tubular organ (tubular portion) such as an esophagus. FIG. 3 shows an endoscope image captured by the endoscope imaging device of FIG. 2, FIG. 4 shows an image processing function by the CPU, and FIG. Shows a flow of a processing procedure for determining whether or not a Barrett's esophagus is passed through a process for generating a development view, and FIG. 6 shows a relationship between the endoscopic image and the generated development view.
FIG. 7 shows a flow of a processing procedure for generating a development view from an endoscopic image, FIG. 8 shows a flow of a processing procedure for detecting the epithelial boundary in FIG. 5, and FIG. 9 shows an average of the calculated epithelial boundary in the Z direction. FIG. 10 shows a development example in which values are displayed, and FIG. 10 shows an example of a histogram in which a reference value serving as a determination criterion is calculated based on a histogram of variance values of samples in the case of epithelial boundary and diagnosis in the case of Barrett's esophagus. Shows an example of display on a monitor displaying an endoscopic image, a developed view, and a determination result, and FIG. 12 shows an endoscopic image and the inner wall of the peripheral portion of the junction between the esophagus and the stomach estimated from the endoscopic image. A three-dimensional shape and a development view generated from the three-dimensional shape are shown.

An endoscope system 1 shown in FIG. 1 includes an endoscope observation apparatus 2 and a personal computer that performs image processing on an endoscope image obtained by the endoscope observation apparatus 2. An endoscopic image processing apparatus (hereinafter simply referred to as an image processing apparatus) 3 as a first embodiment of the image processing apparatus for esophageal mucosa, and a monitor 4 that displays an image processed by the image processing apparatus 3 Composed.
The endoscope observation apparatus 2 includes an endoscope 6 that is inserted into a body cavity, a light source device 7 that supplies illumination light to the endoscope 6, and a camera control that performs signal processing on an imaging unit of the endoscope 6. A unit (abbreviated as CCU) 8 and a monitor 9 for displaying an endoscopic image photographed by the image sensor when a video signal output from the CCU 8 is input.

The endoscope 6 includes an insertion portion 11 that is inserted into a body cavity, and an operation portion 12 that is provided at the rear end of the insertion portion 11. Further, a light guide 13 that transmits illumination light is inserted into the insertion portion 11.
The rear end of the light guide 13 is connected to the light source device 7. Then, the illumination light supplied from the light source device 7 is transferred by the light guide 13 and emitted (transmitted illumination light) is emitted from the distal end surface attached to the illumination window provided at the distal end portion 14 of the insertion portion 11. Illuminate the subject.
An imaging device 17 is provided which includes an objective lens 15 attached to an observation window adjacent to the illumination window, and a charge coupled device (abbreviated as CCD) 16 as a solid-state imaging device disposed at the imaging position of the objective lens 15. is there. The optical image stored on the imaging surface of the CCD 16 is photoelectrically converted by the CCD 16.

The CCD 16 is connected to the CCU 8 via a signal line. When a CCD drive signal is applied from the CCU 8, the CCD 16 outputs a photoelectrically converted image signal. This image signal is signal-processed by a video processing circuit in the CCU 8 and converted into a video signal. This video signal is output to the monitor 9, and an endoscopic image is displayed on the display surface of the monitor 9. This video signal is also input to the image processing device 3.
The image processing device 3 receives a video signal corresponding to an endoscopic image input from the endoscope observation device 2, and outputs an image input unit 21 that performs A / D conversion, and the image input unit 21. It has CPU22 as a central processing unit which performs image processing with respect to image data, and the processing program memory | storage part 23 which memorize | stores the processing program (control program) which performs image processing by this CPU22.

Further, the image processing apparatus 3 is processed by the CPU 22, an image storage unit 24 that stores image data input through the image input unit 21, an information storage unit 25 that stores information processed by the CPU 22, and the like. A hard disk 27 serving as a storage device that stores image data, information, and the like via the storage device interface 26, a display processing unit 28 that performs display processing for displaying image data and the like processed by the CPU 22, and image processing by the user And an input operation unit 29 including a keyboard for inputting data such as parameters and inputting instructions.
The video signal generated by the display processing unit 28 is displayed on the monitor 4, and a processed image subjected to image processing is displayed on the display surface of the monitor 4. The image input unit 21, CPU 22, processing program storage unit 23, image storage unit 24, information storage unit 25, storage device interface 26, display processing unit 28, and input operation unit 29 are connected to each other via a data bus 30. ing.

In the present embodiment, as shown in FIG. 2, for example, an imaging device 17 provided at the distal end portion 14 is inserted into the tubular organ or tubular portion such as the esophagus 31 and the insertion portion 11 of the direct-view endoscope 6 is inserted. Thus, the inner wall mucous membrane of the esophagus 31 is imaged.
FIG. 3 shows an example of an endoscopic image Ia of Barrett's esophagus imaged by this direct-view type endoscope 6. The Barrett's esophagus is one in which the squamous epithelium 32 as the esophageal mucosa is continuously denatured into the gastric mucosa or the columnar epithelium 33 as the Barrett mucosa from the gastroesophageal junction B as the junction between the stomach and the esophagus. A disease called Barrett's esophagus is diagnosed when the Barrett's mucosa occurs 3 cm or more from the normal mucosal boundary and circumferentially with respect to the esophageal lumen cross section.

Including the case defined as described above, the endoscope 6 observes how the degenerated columnar epithelium 33 spreads and the characteristic shape of the epithelial boundary A as the boundary between the columnar epithelium 33 and the squamous epithelium 32. The surgeon diagnoses Barrett's esophagus.
In the endoscopic image Ia of FIG. 3, the tubular part extending from the esophagus 31 to the stomach interior is a gastroesophageal junction B around the darkest portion (not shown), an outer columnar epithelium 33 around the gastroesophageal junction B, The outer epithelial boundary A and the squamous epithelium 32 outside the epithelial boundary A are displayed.

In the present embodiment, a processing for generating an unfolded view by capturing an object of a tubular organ such as the esophagus 31 with the direct-view endoscope 6 and geometrically converting the captured endoscope image Ia. And a developed view of the generated object is displayed on the monitor 4, and the epithelial boundary A is detected from the generated developed view.
Then, based on the characteristic shape of the detected epithelial boundary A, the shape is analyzed, and the analysis result is output.
Specifically, the average value of the epithelial boundary A in the Z direction is calculated, and further, the dispersion value of the epithelial boundary A in the Z direction is calculated to determine whether the epithelial boundary or Barrett's esophagus. Outputs judgment result (analysis result) of Barrett's esophagus.

As shown in FIG. 4, the CPU 22 constituting the image processing apparatus 3 has a geometric image conversion means (function) 22 a that performs geometric conversion, and a development that outputs an image of a development view through the geometric conversion. Figure output means (function) 22b, epithelial boundary detection means (function) 22c for detecting the epithelial boundary from the developed view, and epithelial boundary analysis means (function) 22d for analyzing the shape of the epithelial boundary and outputting the analysis result Have.
Specifically, the epithelial boundary analyzing means 22d includes an epithelial boundary average value calculating means (function) for calculating an average value of the epithelial boundary in the Z-axis direction, and an epithelial boundary for calculating a variance value of the epithelial boundary in the Z direction. Dispersion value calculation means (function) and determination means (function) for determining whether the dispersion value is an epithelial boundary or Barrett's esophagus.

  In the present embodiment, the geometric image conversion means 22a, the developed view output means 22b, the epithelial boundary detection means 22c, and the epithelial analysis means 22d shown in FIG. 4 are realized by software. The CPU 22 reads out the processing program stored (stored) in the storage unit 23, and the CPU 22 executes the processing procedure shown in FIG. 5 according to the processing program. As will be described below, in this embodiment, a developed view Ib is generated from an endoscopic image Ia obtained by imaging the gastroesophageal joint peripheral portion of the esophagus with a direct-view type endoscope 6, and the epithelial boundary is generated from the developed view Ib. In addition, the shape of the Barrett's esophagus with respect to the shape of the epithelial boundary A, that is, when each point of the epithelial boundary A is jagged in the luminal direction, Whether or not the feature amount is present is determined by calculating a variance value at each epithelial boundary point.

Then, a typical Barrett esophagus is objectively determined when each point of the epithelial boundary A in the captured endoscopic image Ia has a complicated shape, such as jagged in the luminal direction. I have to.
Next, the operation of this embodiment will be described with reference to FIG.
When the operation of the image processing apparatus 3 starts, the CPU 22 reads the processing program in the processing program storage unit 23 and starts processing according to the processing program. In the first step S1, the CPU 22 acquires image data of the endoscopic image Ia input from the CCU 8 of the endoscope observation apparatus 2 via the image input unit 21.

In the next step S2, the CPU 22 performs preprocessing such as distortion correction (see, for example, JP-A-8-256295) and noise removal on the acquired image data. ) To generate a development view Ib shown in FIG. 6B from the endoscopic image Ia shown in FIG.
FIG. 7 shows a process for generating the development diagram Ib.
As shown in FIG. 7, in the first step S11, the position of the darkest part in the endoscopic image Ia is detected, and the center of gravity position of the detected darkest part is set as the center position of the coordinates of the endoscopic image Ia.
In this embodiment, a development view is generated around the position of the darkest part in the endoscopic image Ia. As a method for detecting the darkest part, the endoscopic image Ia is divided into a plurality of areas, the average luminance of the divided areas is calculated, and the area having the minimum average luminance is obtained as the position of the darkest part.

As shown in FIG. 6, a two-dimensional orthogonal coordinate system of the endoscopic image Ia is set to XY, and a development view Ib is geometrically converted from the endoscopic image Ia to the polar coordinate system to be θ-Z. Generate. The coordinate position in the coordinate system XY is indicated by x and y. Further, the coordinate position in the polar coordinate system θ-Z is represented by θ representing the position in the circumferential direction, and z the distance position from the center.
In FIG. 6, in order to facilitate understanding of the relationship when the developed view Ib is generated from the endoscopic image Ia, the epithelial boundary A between the squamous epithelium 32 and the columnar epithelium 33 in the endoscopic image Ia, the gastroesophageal junction When B is a developed view Ib, the shape of the display is indicated by an arrow in association with the arrow. Here, the scale is shown with θ being 0 degrees, 45 degrees, 90 degrees,. The position of the image display frame is indicated by Q ₀ , Q ₄₅ , and Q ₉₀ .
In the next step S12 and step S13, the initial value of the coordinates S (θ, z) of the developed view Ib is set. That is, in step S12, the CPU 22 sets θ = 0, and sets z = 0 in step S13.

The coordinate position of the endoscopic image Ia set in the next step S14 and corresponding to the coordinate S (θ, z) of the developed view Ib is obtained by the following equation (1).
x = z sin θ, y = z cos θ (1)
It is determined whether the coordinates P (x, y) calculated in step S15 exist in the endoscopic image Ia.
And when it exists in the endoscopic image Ia, it progresses to step S16. Since the position of the coordinate P (x, y) of the endoscopic image obtained from the expression (1) may exist between pixels, the coordinate P is obtained using processing such as linear interpolation in step S16. The luminance value of (x, y) is calculated.
As the luminance value, when color imaging is performed, the luminance value of each color signal is calculated.

In step S17, the luminance value obtained in step S16 is set as the luminance value of the coordinate S (θ, z) in the development view. Next, the process proceeds to step S18, the value of z in the developed view Ib is changed (for example, z increment Δz = 1), and the process returns to step S14.
On the other hand, if the coordinate P (x, y) calculated in step S15 does not exist in the endoscopic image Ia, the process proceeds to step S19, and the value of θ in the developed view Ib is changed (for example, an increment of θ Δθ = π). / 180 or 1 °).
In the next step S20, if [theta] is smaller than 2 [pi] (360 [deg.]), The process returns to step S13 to continue the process of generating the development diagram. On the other hand, if θ is equal to or greater than 2π, it is determined that the development view Ib has been generated, and the process proceeds to step S11. The endoscopic image Ia and the development view Ib are output to the monitor 4, and the processing for generating the development view is completed. Then, the process proceeds to the process of detecting the epithelial boundary in step S4 of FIG. 5 using the generated developed view Ib.

In order to detect the epithelial boundary as a boundary between the squamous epithelium and the columnar epithelium shown in step S4 of FIG. 5 (acquire a coordinate point sequence), for example, edge detection is used as shown in FIG. Since the squamous epithelium has a white color tone, and the columnar epithelium has a red color tone, the epithelial boundary can be detected by a specific color tone in R, G, and B.
As shown in FIG. 8, in step S31, the CPU 22 acquires the image data of the development view Ib stored in the image storage unit 24 or the like.
In the next step S32, the CPU 22 performs the known edge detection process described above as the contour extraction filtering process on the image data to generate an edge detection image.

In the next step S33, the CPU 22 performs binarization processing on the edge detection image, and further performs thinning processing to acquire a boundary.
In the next step S34, the CPU 22 performs a tracing process on the acquired boundary to acquire a coordinate point sequence along the boundary, that is, a point sequence of the epithelial boundary.
After acquiring the epithelial boundary point sequence in this way, the CPU 22 calculates the average value <Z _A > in the Z direction (which is the esophageal lumen direction) of the epithelial boundary as shown in step S5 of FIG.
That is, the average value <Z _A > in the Z direction of the detected epithelial boundary is obtained by the following equation.

However, Z _Aθi is the Z value at θi of the epithelial boundary in the developed view obtained by Expression (1), and Expression (2) takes an average with the sample value N for one rotation in the circumferential direction. (In this case, the resolution of the developed view of FIG. 6 is determined by the fineness of sampling of θi and Z (distance from the center O).

FIG. 9 shows a diagram in which an average value <Z _A > in the Z direction of the epithelial boundary is displayed on the developed view Ib. After calculating the average value <Z _A > of Equation (2), a variance value σ _A representing the degree of variation in the coordinate position of the epithelial boundary in the Z direction (unevenness of the boundary) is calculated by the following equation in Step S6. Ask.

In the next step S7, by comparing the equation (3) variance sigma _A and the reference value sigma _th calculated in this variance sigma _A it is determined whether the reference value sigma _th smaller.
If the variance value σ _A is smaller than the reference value σ _th, it is determined as an epithelial boundary as shown in step S8. Conversely, if the variance value σ _A is larger than the reference value σ _th , the epithelium as shown in step S9. It is determined that the boundary has a complicated shape, and the captured image is determined as Barrett's esophagus.
In this case, the reference value σ _th used for the determination with respect to the variance value σ _A obtained from the development diagram Ib uses a sample of the epithelial boundary and a sample in the case of Barrett's esophagus that have been diagnosed in advance as shown in FIG. A histogram of the variance value σ obtained in this way is used.

Figure 10 shows a histogram _{D nor} the distribution of the variance value sigma _nor obtained when epithelial border, and a histogram _{D bar} distribution of the resulting dispersion value sigma _bar in the case of Barrett's esophagus. In the case of Barrett's esophagus, a sample having a complicated boundary shape such as a jagged epithelial boundary is used.
As shown in FIG. 10, from two histograms D _nor and D _bar , for example, a position σ _th where the two histograms intersect is set as a reference value for evaluating the Barrett esophagus.
Then, as described above, the variance value σ _A obtained by the equation (3) is compared with the reference value σ _th to determine whether it is an epithelial boundary or Barrett's esophagus.
Information on the determination result (analysis result) in step S8 or step S9 is output to the display device such as the monitor 4 as shown in step S10, and this process is terminated.

FIG. 11 shows a display example in which the endoscope image Ia, the development view Ib, and the information of the determination result are displayed on the monitor 4. In the example of FIG. 11, as an example when the variance value σ _A is equal to or greater than the reference value σ _th, the determination result determined as “Barrett's esophagus” is displayed on the determination display unit 41.
Here, the Barrett's esophagus is displayed, but “Barrett's esophagus is highly likely” may be displayed. Further, instead of obtaining the variance value σ _A , a standard deviation value may be calculated and compared with a reference value, and a comparison result may be used as an analysis result.
Further, in the present example and the following examples, the comparison result obtained by comparing the evaluation value for the feature amount of the Barrett's esophagus to be determined, such as the calculated dispersion value σ _A , with a single reference value (σ _th etc.) Although it is an analysis result, the calculated evaluation value is compared with a plurality of reference values, and an intermediate state (one or a plurality of intermediate states) between a state where the possibility of Barrett's esophagus is high and a state close to normal The analysis result or the like may be output.

For example, when it is determined that the squamous epithelium as the esophageal mucosa is slightly degenerated compared to the case of the epithelial boundary, there is a possibility that the columnar epithelium (Barrett mucosa) is slightly generated, If it is determined that the condition is more degenerated, it may be notified by displaying a determination result indicating that the symptom in which the columnar epithelium (Barrett mucosa) is generated is in an advanced state.
The surgeon can make an efficient diagnosis by referring to the information of the determination result. In addition, the surgeon refers to the developed view Ib displayed together with the endoscopic image Ia displayed on the monitor 4, so that the shape of the epithelial boundary A in the circumferential direction (from the case of only the endoscopic image Ia) B) it is easier to grasp, and therefore diagnosis and the like can be carried out efficiently.

In the above description, the development view Ib is generated from the two-dimensional endoscopic image Ia by the geometric transformation by the process shown in FIG. 7, but the endoscopic view shown in FIG. A developed view Ib shown in FIG. 12C is generated from the mirror image Ia by estimating the three-dimensional shape of the inner wall surface around the gastroesophageal junction as shown in FIG. 12B. You may make it do. When the solid shape is estimated as shown in FIG. 12B, a three-dimensional coordinate system (X _L , Y _L , Z) in which the center line of the solid shape is further estimated and the coordinate axis Z _L is set in the direction of the center line. _L ), each point of the endoscopic image Ia is converted.
Further, each point of the three-dimensional shape in the three-dimensional coordinate system (X _L , Y _L , Z _L ) is projected onto the surface of a cylindrical body centered on the same Z _L close to this three-dimensional shape. The coordinate system of the surface of this cylinder is defined as θ _L -Z _L. Then, each point of the three-dimensional coordinate system (X _L , Y _L , Z _L ) or each point projected on the surface of the cylindrical body is converted into a coordinate system of θ _L -Z _L.

In FIG. 12B, a point P ′ on the epithelial boundary A is shown, and the point P ′ on the epithelial boundary A is indicated by an angle θ _L with the Y _L axis.
As shown in FIG. 12 (B), each point projected on the surface of the cylinder is converted into a coordinate system of θ _L -Z _L , and then opened by a value where θ _L is 0, and the development shown in FIG. 12 (C). Each position of the endoscopic image Ia is displayed according to FIG. Ib. The details of this processing are described in the specification of Japanese Patent Application No. 2005-1842.
Further, another development drawing generation means described in this specification may be adopted to generate a development drawing (also in the case of generating a development drawing by two-dimensional conversion, Japanese Patent Application No. 2004-2004). It is also possible to adopt a development drawing generation means described in the specification of No. 378011 to generate a development drawing and perform the above-described processing using the development drawing.

Figure 12 In this case, as shown in (C), Z _L because is a position distant from the large value side imaging means, epidermal boundary A below the gastroesophageal junction B corresponding to the distant position (inside) The shape of is displayed.
As described above, also in the development view Ib shown in FIG. 12C, the average of the epithelial boundary A in the Z direction (ZL direction in FIG. 12C) is obtained by performing the processing after step S4 in FIG. The value <Z _A > and the variance value σ _A are calculated, and the variance value σ _A is compared with the reference value σ _th to determine whether it is an epithelial boundary or a Barrett esophagus whose properties have changed. it can. Note that the resolution of the developed view of FIG. 12C is determined by the resolution of the image when the three-dimensional shape to be diagnosed is estimated.

As described above, in this embodiment, the development view Ib of the endoscopic image of the esophagus imaged by the direct-view type endoscope 6 is generated, and the shape of the epithelial boundary A in the development view Ib is quantitatively determined. By doing so, it can be easily and objectively determined whether or not it is Barrett's esophagus.
Therefore, the surgeon can efficiently diagnose whether or not it is Barrett's esophagus by using this determination result.
Further, in the present embodiment, the development view Ib is generated, so that the shape distribution of the epithelial boundary in the lumen direction or the circumferential direction orthogonal to the lumen can be grasped more than in the case of only the endoscopic image Ia. Therefore, diagnosis is also easier to perform.

Next, a second embodiment of the present invention will be described with reference to FIGS. The Barrett's esophagus is a case where the epithelial boundary is applied to a complicated shape as described in Example 1, but the columnar epithelium (Barrett mucosa) spreads in a tongue shape. May be formed.
In this case, in the first embodiment, it may be difficult to output an appropriate analysis result for this shape. Therefore, an object of the present embodiment is to output an appropriate analysis result (determination result) even when the columnar epithelium spreads in a tongue shape.
Therefore, in this embodiment, as described below, the average value <Z _A > in the Z direction of the epithelial boundary A in the first embodiment is calculated to deal with the variation (complexity) of the epithelial boundary A in the Z direction. Instead of calculating the variance value σ _A to be calculated, the maximum value and the minimum value in the Z direction of the epithelial boundary A are obtained, and the absolute value of these difference values is compared with the reference value. Determine if it is an esophagus.

FIG. 13 shows an endoscopic image Ia obtained by imaging the vicinity of the gastroesophageal junction, and a developed view Ib generated two-dimensionally or three-dimensionally from the endoscopic image Ia.
In this embodiment, a process as shown in FIG. 14 is performed. The processing in FIG. 14 is the same as the processing in FIG. 5 from step S1 to step S4. After detecting the epithelial boundary A in step S4, in this embodiment, the maximum value Zmax and the minimum value Zmin in the Z direction of the epithelial boundary A are calculated as shown in step S41.
Then, at the next step S42, the absolute value _{D A} of the difference values of the maximum value Zmax and the minimum value Zmin.
Determines at next step S43, the absolute value D _A of the difference value is compared with a reference value D _th, the absolute value D _A of the difference value or reference value D _th is smaller than or not.

This reference value D _th is obtained by obtaining a histogram of absolute values of difference values in a normal sample whose diagnosis has been confirmed in advance and a Barrett's esophagus sample in which the epithelial boundary A extends in a tongue shape. It is calculated as a threshold suitable for performing.
When the absolute value D _A of the difference value is the reference value D _th smaller than determines that the epithelium boundary as shown in step S8, step S9. If the absolute value D _A of the difference value in the opposite is equal to or larger than the reference value D _th It is determined as Barrett's esophagus as shown in. Then, the result of determination in step S8 or step S9 is displayed on a display device such as the monitor 4 in step S10, and the determination result is output, and this process is terminated.
According to the present embodiment, even in the case of Barrett's esophagus where the epithelial boundary A spreads out in a tongue shape, it is possible to appropriately analyze the characteristics and make an accurate determination.

Next, Embodiment 3 of the present invention will be described with reference to FIGS. The object of the present embodiment is to appropriately determine whether or not it is a Barrett esophagus from the characteristic shape of the epithelial boundary even when the shape of the lumen of the esophagus is deformed from, for example, a circular tube shape.
In the determination method of Example 1, assuming that the esophagus is close to a circular tube shape, an average value <Z _A > of the epithelial boundary A in the Z direction is calculated, and each position of the epithelial boundary A in the Z direction is calculated as the average. _A variation (dispersion value) from the value <Z _A > was calculated to determine whether or not it was Barrett's esophagus. In this method, when the shape of the lumen of the esophagus is deformed from the circular tube shape due to, for example, the influence of the pulsation of the heart or the like, the determination result is easily affected.
In the present embodiment, in order to reduce the influence, a development view Ib is generated from the endoscopic image Ia shown in FIG. 15, and for example, the gastroesophageal junction B is detected on the development view Ib. It is determined from the dispersion value of the distance from the part B to the epithelial boundary A whether or not it is a Barrett esophagus.

That is, when the lumen shape of the esophagus is deformed, the gastroesophageal junction B may be deformed in substantially the same manner. Therefore, by using the distance information from the gastroesophageal junction B to the epithelial boundary, The determination is made almost without being influenced by the cavity shape. In this case, instead of the gastroesophageal junction B, a dark part or a cardia part leading to the stomach may be detected and used.
As a method for detecting the gastroesophageal junction B, for example, the end point of the fence-like blood vessel running in the lumen direction (that is, the Z direction) of the esophageal mucosa is detected, and the end point is connected to detect the gastroesophageal junction B. be able to.
The processing method of this embodiment is as shown in FIG. In FIG. 16, steps S1 to S4 are the same as those in FIG. 5 or FIG. Following step S4, the process of detecting the gastroesophageal junction B in step S51 is performed to detect the gastroesophageal junction B.

In the next step S52, an average value <Z _AB > and a variance value σ _AB of the distance Z _Aθi -Z _{Bθi in} the Z direction between the epithelial boundary A and the gastroesophageal junction Z are calculated.
Here, the average value <Z _AB > and the variance value σ _AB can be obtained by the following equations.

Then, in the next step S53, it is determined whether the variance value σ _AB is the epithelial boundary or the Barrett's esophagus whose properties have changed by comparing with the reference variance value σ _ABth used for the previously obtained determination.
That is, if the variance value σ _AB is smaller than the reference variance value σ _ABth in step S53, it is determined as an epithelial boundary as shown in step S8. Conversely, the variance value σ _AB is the reference variance value σ. If it is greater than or equal to _ABth, it is determined to be Barrett's esophagus as shown in step S9.

Then, as shown in step S10, the determination result is output by, for example, displaying the determination result in step S8 or step S9 on the monitor 4, and the process is terminated.
According to the present embodiment, the average value <Z _AB > of the distance Z _Aθi −Z _Bθi to the epithelial boundary A is calculated based on the gastroesophageal junction B, and the variation of the distance Z _Aθi −Z _Bθi is evaluated. Thus, since it is determined whether the epithelial boundary A is the epithelial boundary or Barrett's esophagus, it can be appropriately determined even when the lumen of the esophagus is deformed.
As described above, instead of using the gastroesophageal joint B as a reference, the determination may be made using a dark part or a cardia part leading to the stomach as a reference. For example, in order to detect a dark part (darkest part), binarization is performed with a threshold value set for detecting the dark part, the boundary is detected by edge extraction, and the boundary of the dark part is detected by further thinning the line. it can.
In this case, the calculation can be performed by a simpler process than when the gastroesophageal junction B is detected.

Next, a fourth embodiment of the present invention will be described with reference to FIGS. In this embodiment, in the case of Barrett's esophagus, paying attention to the feature that the epithelial boundary A has a complicated shape, the line length of the epithelial boundary A is calculated to determine whether the feature is the epithelial boundary or Barrett's esophagus. Do.
FIG. 17 shows a developed view Ib generated from an endoscopic image as in the above-described embodiments. Since this development view Ib is a discrete image, the curve of the epithelial boundary A is actually represented by a broken line. Accordingly, distances L _i-1 and _i between two adjacent points A _i-1 (θ _i-1 , Z _i-1 ) and A _i (θ _i , Z _i ) are calculated, and the distance L _i-1 is calculated. , by obtaining the sum of _i, determine the length L _a of the epithelium boundary a. That means

  Here, i is a value from 1 to N when 360 degrees (2π) is divided into N equal parts.

It is determined whether the epithelium boundary or Barrett's esophagus by the procedure shown in FIG. 18 by using the length L _A of the thus calculated epithelium boundary A. Steps S1 to S3 in FIG. 18 are omitted because they are the same as those in FIG.
After detecting the epithelial boundary A in step S4 as shown in FIG. 18, the length LA of the epithelial boundary _A is calculated as shown in step S61.
In the next step S62, the length L _A of the calculated epithelial boundary A, compared with the length L _th epithelial border as a reference to be used for determining using a number of samples beforehand diagnosis is confirmed. When the length L _A is the reference length L _th smaller epithelial border epithelial border A determines that the epithelium boundary as shown in step S8, the length L _A reference epithelial boundary A in the opposite in the case of more than the length L _th epithelial boundary determines that a Barrett's esophagus as shown in step S9.

Then, as shown in step S10, the determination result is output by, for example, displaying the determination result in step S8 or step S9 on the monitor 4, and the process is terminated.
According to the present example, in the case of Barrett's esophagus, there are many cases where the epithelial boundary A has a complicated shape, and an accurate determination can be made for such cases.
In the case of calculating the length L _A of the epithelium boundary A, so as to represent the curve of the epithelium boundary A in approximate expression may be calculated line length by the approximate expression. In this way, a more accurate line length can be calculated, and determination based on the line length can also be performed with higher accuracy.
Note that the distance L _i-1 , _i between two adjacent points A _i-1 (θ _i-1 , Z _i-1 ), A _i (θ _i , Z _i ) is the entire circumference (360 degrees) in the circumferential direction. minute and from the sum value to calculate the length L _a of the epithelial border a but not necessarily limited to the entire circumference. The same applies to other embodiments.

Next, Embodiment 5 of the present invention will be described with reference to FIGS. In this embodiment, the sum of the absolute values of the angles formed by adjacent vectors in the curve of the epithelial boundary in the circumferential direction is obtained, and it is determined whether the epithelial boundary or Barrett's esophagus is based on this sum.
Also in the present embodiment, a development view Ib is generated from the endoscopic image Ia as shown in FIG. Then, the epithelial boundary A is detected from the developed view Ib. Since the developed view Ib is a discrete image, the epithelial boundary A is represented by a polygonal line.
Three adjacent points on the epithelial boundary A are extracted, and an angle formed by two line segments each connecting the two adjacent points or an angle formed by two vectors is obtained.
For example, an enlarged view of a circle C in FIG. 19 is shown in FIG. 19B. Vectors V _i and V _{i + 1} at three points A _i−1 , A _i , and A _{i + 1} in this figure are enlarged and shown in FIG. C).

Then, an angle φ _θi formed by both vectors V _i and V _{i + 1} is calculated.
Then, the total value φ _A of the absolute value of the angle φ _{θi in} that case is calculated while moving the three points A _i−1 , A _i , A _{i + 1} to the left side so that i is increased one by one from 0 _. .
in this case,

It becomes.
The method for determining whether the epithelial boundary or Barrett's esophagus in this embodiment is as shown in FIG. Also in FIG. 20, since the first step S1 to step S3 are the same as those in FIG. 5, the steps are omitted.

After detecting the epithelial boundary A as shown in step S4, in the next step S71, as described in FIG. 19, two vectors V _i and V _{i + 1} are sequentially set from three adjacent points on the development view Ib. Then, an angle φ _θi formed by them is calculated. Then, at the next step S72, the calculating a sum phi _A of the absolute values of these corners phi _.theta.i.
In the next step S73, the calculated sum total φ _A is compared with the sum value φ _th for the reference epithelial boundary obtained using a number of samples whose diagnosis has been confirmed in advance. If the calculated total sum φ _A is smaller than the reference total value φ _th, it is determined as an epithelial boundary as shown in step S8. Conversely, if the total sum φ _A is greater than or equal to the reference total value φ _th , It determines with it being a Barrett esophagus as shown to step S9.

Then, as shown in step S10, the determination result is output by, for example, displaying the determination result in step S8 or step S9 on the monitor 4, and the process is terminated.
According to the present example, in the case of Barrett's esophagus, there are many cases where the epithelial boundary A has a complicated shape, and an accurate determination can be made for such cases.

Next, Embodiment 6 of the present invention will be described with reference to FIG. In this embodiment, the number of inflection points of the epithelial boundary is calculated, and it is determined whether the epithelial boundary or Barrett's esophagus is large or small. As described above, in the case of Barrett's esophagus, a complicated shape such as a jagged shape is often accompanied. For example, as shown in FIG. 19A, the epithelial boundary A often has an uneven shape in the case of Barrett's esophagus.
In this embodiment, as a specific example of calculating (evaluating) the number of bending points, for example, the number of inflection points corresponding to the maximum value and the minimum value of the epithelial boundary A in the development view is calculated. Alternatively, a line segment or vector connecting two adjacent points on the epithelial boundary A in the developed view is in a state where the angle between the horizontal direction and the vector is more than 0 degrees, or less than 0 degrees (positive to negative), or less than 0 degrees to 0 degrees Count the number of points above (from negative to positive). This is shown in FIG.

As shown in FIG. 21, the slope ΔZ _i / Δθ _i = Ψ _i of the vector V _i of the two points A _i-1 and A _i on the epithelial boundary A changes from positive to negative, and the inflection changes from negative to positive. Calculate the number of points. In FIG. 21, since the slope of the vector V _{i + 2} becomes negative, the point A _{i + 2} is determined as the inflection point.
Then, the inflection points are detected in this way from θ _i to i = 1 to N, and the sum of the inflection points is calculated. Then, it is determined whether the epithelial boundary or Barrett's esophagus is based on the sum of the inflection points.
The method for determining whether the epithelial boundary or Barrett's esophagus in this embodiment is as shown in FIG. Also in FIG. 22, since the first step S1 to step S3 are the same as those in FIG. 5, the steps are omitted.

After detecting the epithelium boundary A, as shown in step S4, in the next step S81, the variable parameter i of theta, it initializes the number F _A bending point. That is, i = 1 and F _A = 0. In the next step S82, the calculated inclination [psi _i the vector _{V i} as shown in FIG. 21. Then, in the next step S83, it is determined whether this slope Ψ _i corresponds to an inflection point. This determination, in the case corresponding to the inflection point, after increasing one value of the number F _A, as shown in step S84, the process proceeds to the next step S85.
On the other hand, if it does not correspond to the inflection point, the process proceeds to step S85. In step S85, it is determined whether the variable parameter i is equal to the division number N of θ. If the variable parameter i is not equal to N, the number of variable parameters i is increased by one as shown in step S86, and the process returns to step S82.

In this way, the processing from step S82 to step S86 is repeated until i = N.
When i = N, it is determined whether the number F _A is smaller than the reference number F _th as shown in step S87. The reference number F _th is calculated from a large number of samples whose diagnosis is confirmed.
If the number F _A is smaller than the reference number F _th , the epithelial boundary is determined as shown in step S8. Conversely, if the number F _A is greater than or equal to the reference number F _th , the process proceeds to step S9. As shown, it is determined to be Barrett's esophagus.
Then, as shown in step S10, the determination result is output by, for example, displaying the determination result in step S8 or step S9 on the monitor 4, and the process is terminated.

According to the present example, in the case of Barrett's esophagus, there are many cases where the epithelial boundary A has a complicated shape, and an accurate determination can be made for such cases.
In the present embodiment, an example has been described in which the number of inflection points F _A is calculated, and the calculated number is compared to determine whether it is an epithelial boundary or Barrett's esophagus. As described in JP-A-360319, the number of inflection points may be calculated and compared to determine whether it is an epithelial boundary or Barrett's esophagus.
This determination method may be performed in accordance with the processing of steps S4 to S14 in FIG. 4 of Japanese Patent Application No. 2004-360319. In this case, it is performed on image data obtained from an endoscopic image.

In the above description, the number of inflection points F _A or the number of inflection points is calculated and compared with a reference value to determine whether the epithelial boundary or Barrett's esophagus is used. The two slopes Ψ _i and Ψ _{i + 1} are compared, the number of points at which the value changes (for example, the point at which the subtraction value changes) is counted, and the sum is compared with the reference value to make a determination. You may make it do.
For example, in the processing procedure of FIG. 5, after the development view is generated in step S3, the epithelial boundary is detected in the next step S4. However, after the epithelial boundary is detected first, the development view is generated. You may do it.
Further, in each of the above-described embodiments, a case has been described in which it is quantitatively determined whether or not it is Barrett's esophagus as a disease around the gastroesophageal junction in the esophagus as the upper gastrointestinal tract. It can also be applied to.

  For example, a developed view is generated from an image captured by inserting the endoscope 6 into the large intestine, and the boundary of the lesioned mucous membrane having a color tone different from that of the normal site is detected from the developed view. For example, an analysis result may be obtained by analyzing the shape of the boundary. In short, the present invention can be widely applied as an apparatus and method for generating a development view from an endoscopic image obtained by imaging a tubular part inside the body and analyzing a lesioned part or the like from the development view. In this case, by using the analysis result together with the development view, it is easy to grasp the lesioned part and the like, and diagnosis and the like can be performed efficiently. Note that embodiments configured by partially combining the above-described embodiments also belong to the present invention.

[Appendix]
1. In Claim 1, the said analysis means calculates the difference value of the maximum value with respect to the luminal direction of the said esophagus in the said epithelial boundary, and makes a comparison result with a predetermined value an analysis result.
2. 2. The analysis unit according to claim 1, wherein the analysis unit calculates a variation amount of a distance from a reference site such as a joint between the stomach and the esophagus to the epithelial boundary, and compares the calculated variation amount with a reference value to obtain an analysis result.
3. In Claim 1, the said analysis means calculates the sum total value of the distance between the two adjacent points on the said epithelial boundary, and makes the comparison result which compared the calculated sum total value with a reference value an analysis result.

4). In Claim 1, the said analysis means calculates the sum total value of the absolute value of the angle | corner which two vectors (line segment) which ties two adjacent points in the three adjacent points on the said epithelial boundary, and calculated the total value The comparison result of comparing with the reference value is taken as the analysis result.
5. 2. The analysis unit according to claim 1, wherein the analysis means calculates the number of bending points when the shape of each point on the epithelial boundary is uneven with respect to the luminal direction of the esophagus, and uses the calculated number of bending points as a reference value. The comparison result compared is taken as the analysis result.
6). 2. The comparison result according to claim 1, wherein the analysis means calculates the number of extreme values at which the shape of the epithelial boundary is maximum and minimum with respect to the luminal direction of the esophagus, and compares the calculated number of extreme values with a reference value. Is the analysis result.

7). A development view generation step for generating a development view for an image of the inside of the esophagus;
An epithelial boundary detection step for detecting an epithelial boundary as a boundary between the squamous epithelium as the esophageal mucosa from the developed view and the columnar epithelium in which the squamous epithelium is denatured into the gastric mucosa,
An analysis step for calculating an analysis result corresponding to a predetermined feature amount for the detected epithelial boundary;
An image processing method for esophageal mucosa, comprising:
8). In Supplementary Note 7, the analysis step calculates an evaluation value of a feature value of an epithelial boundary in the case of Barrett's esophagus as the predetermined feature value, and compares the evaluation value with a reference value serving as a reference to calculate an analysis result It is characterized by that.

9. In Supplementary Note 7, the analyzing step includes an average value calculating step for calculating an average value of the epithelial boundary and a variation amount calculating step for calculating a variation amount from the average value of the epithelial boundary.

10. In Supplementary Note 7, the analysis step includes a step of calculating a difference value between a maximum value and a minimum value with respect to a lumen direction of the esophagus at the epithelial boundary.
11. In Supplementary Note 7, the analysis step includes a step of calculating a variation amount of a distance from a reference site such as a junction between the stomach and the esophagus to the epithelial boundary.
12 In Supplementary Note 7, the analysis step includes a step of calculating a total value of distances between two adjacent points on the epithelial boundary.

13. In Supplementary Note 7, the analysis step includes a step of calculating a sum of absolute values of angles formed by two vectors (line segments) connecting two adjacent points on three adjacent points on the epithelial boundary.
14 In Supplementary Note 7, the analysis step includes a step of calculating the number of bending points when the shape of each point on the epithelial boundary becomes uneven with respect to the luminal direction of the esophagus.

15. In Supplementary Note 7, the analysis step includes a step of calculating the number of extreme values at which the shape of the epithelial boundary becomes maximum and minimum with respect to the luminal direction of the esophagus.
16. An endoscope provided with imaging means for imaging an image of a tubular organ in a body cavity;
A developed view generating means for generating a developed view of the tubular organ using the captured at least one image;
An epithelial boundary detecting means for detecting an epithelial boundary from the developed view;
Analyzing the shape of the boundary detected by the epithelial boundary detection means, and an analysis means for outputting an analysis result;
An endoscope apparatus characterized by comprising:

17. Development view generating means for generating a development view for an image of the inside of the esophagus;
An epithelial boundary detecting means for detecting an epithelial boundary as a boundary between the squamous epithelium as the esophageal mucosa from the image or the developed view and the columnar epithelium in which the squamous epithelium is denatured into the gastric mucosa,
An analysis means for calculating an analysis result corresponding to a predetermined feature amount with respect to the detected epithelial boundary;
An image processing apparatus for esophageal mucosa, comprising:
18. A development view generating means for generating a development view from an endoscopic image obtained by imaging a tubular part in a living body;
Analysis means for evaluating (analyzing) the characteristic amount of the mucosal tissue to be determined from the developed view and outputting the result;
A medical image processing apparatus comprising:

  A development view is generated from an endoscopic image obtained by imaging the inside of the esophagus to detect the epithelial boundary between the squamous epithelium as the esophageal mucosa and the degenerated columnar epithelium, and the Barrett esophagus at the epithelial boundary By analyzing the feature value of each disease and calculating the analysis evaluation value for the feature value, and comparing the analysis evaluation value with the reference value, an objective analysis result can be obtained. Diagnosis and the like can be carried out efficiently by using together with a development view in which the shape of the shape can be easily grasped.

FIG. 1 is a block diagram illustrating a configuration of an endoscope system including the first embodiment of the present invention. FIG. 2 is a view showing a state where an image is taken by an endoscope inserted into a tubular organ (tubular portion) such as an esophagus. FIG. 3 is a diagram illustrating an endoscopic image captured by the endoscope imaging apparatus of FIG. 2. FIG. 4 is a diagram showing an image processing function by the CPU. FIG. 5 is a flowchart showing a processing procedure for determining whether or not it is a Barrett esophagus through processing for generating a development view. FIG. 6 is a diagram illustrating a relationship between an endoscopic image and a generated development view. FIG. 7 is a flowchart of a processing procedure for generating a development view from an endoscopic image. FIG. 8 is a flowchart of a processing procedure for detecting an epithelial boundary in FIG. FIG. 9 is a development view showing the average value of the calculated epithelial boundary in the Z direction. FIG. 10 is a diagram illustrating an example of a histogram in which a reference value serving as a determination reference is calculated based on a histogram of variance values of samples in the case of an epithelial boundary whose diagnosis has been confirmed in advance and in the case of Barrett's esophagus. FIG. 11 is a diagram illustrating a display example on a monitor displaying an endoscopic image, a development view, and a determination result. FIG. 12 is a diagram showing an endoscopic image, a three-dimensional shape of the inner wall of the peripheral portion of the gastroesophageal junction estimated from the endoscopic image, and a development view generated from the three-dimensional shape. FIG. 13 is a diagram showing an endoscopic image and a development view according to the second embodiment of the present invention. FIG. 14 is a flowchart showing a processing procedure for determining whether or not a Barrett's esophagus is passed through processing for generating a development view. FIG. 15 is a diagram illustrating an endoscopic image and a development view according to the third embodiment of the present invention. FIG. 16 is a flowchart showing a processing procedure for determining whether or not it is a Barrett esophagus through processing for generating a development view. FIG. 17 is an explanatory diagram for calculating the distance between two adjacent points on the epithelial boundary in the developed view. FIG. 18 is a flowchart showing a processing procedure for calculating the sum of the distances between two adjacent points of the epithelial boundary and determining whether or not it is Barrett's esophagus. FIG. 19 is an explanatory diagram showing a state in which an angle formed by two adjacent vectors generated from three adjacent points on the epithelial boundary of a developed view generated from an endoscopic image is obtained. FIG. 20 is a flowchart showing a processing procedure for determining whether or not a Barrett's esophagus is obtained by calculating a sum of angles formed by two adjacent vectors. FIG. 21 is an explanatory diagram showing a state in which the inclination of a vector connecting two adjacent points on the epithelial boundary of the development view is calculated. FIG. 22 is a flowchart showing a processing procedure for calculating the total number of inflection points at which the vector inclination changes from positive to negative and from negative to positive, and determining whether the vector is a Barrett esophagus from the total number.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Endoscope system 2 ... Endoscope observation apparatus 3 ... Image processing apparatus 4 ... Monitor 6 ... Endoscope 7 ... Light source device 8 ... CCU
11 ... Insertion part 14 ... Tip part 16 ... CCD
17 ... Imaging device 21 ... Image input unit 22 ... CPU
22a ... Geometric image conversion means 22b ... Development view output means 22c ... Epithelial boundary detection means 22d ... Epithelial boundary analysis means 23 ... Processing program storage section 24 ... Image storage section 25 ... Analysis information storage section 27 ... Hard disk 28 ... Display processing 29: Input operation unit 31 ... Esophageal A ... Epithelial boundary B ... Gastroesophageal junction Agent Attorney Susumu Ito

Claims (11)

Development view generating means for generating a development view for an image of the inside of the esophagus;
Epithelial boundary detection means for detecting an epithelial boundary as a boundary between the squamous epithelium as the esophageal mucosa and the columnar epithelium of the gastric mucosa based on the color tone change of the pixel value on the developed view or the edge state of the image ,
Feature amount calculating means for calculating feature amounts relating to position information of each pixel constituting the epithelial boundary detected by the epithelial boundary detecting means ;
Analysis means for outputting an analysis result based on the feature quantity calculated by the feature quantity calculation means ;
An image processing apparatus for esophageal mucosa, comprising:   Epithelial boundary position average value calculating means for calculating an average value of the positions of pixels constituting the epithelial boundary in the luminal direction of the esophagus;
  An epithelial boundary for calculating a variance value of the positions of the pixels constituting the epithelial boundary in the luminal direction of the esophagus with respect to the average value of the positions of the pixels constituting the epithelial boundary calculated by the epithelial boundary position average value calculating means Position variance calculation means;
  Further comprising
  The feature amount calculating unit calculates the feature amount based on the variance value calculated by the epithelial boundary position variance value calculating unit.
  The image processing apparatus for esophageal mucosa according to claim 1.   Further comprising an epithelial boundary position standard deviation value calculating means for calculating a standard deviation value of a position of a pixel constituting the epithelial boundary in the lumen direction of the esophagus
  The feature amount calculating means calculates the feature amount based on the standard deviation value calculated by the epithelial boundary position standard deviation value calculating means.
  The image processing apparatus for esophageal mucosa according to claim 1.   The feature amount calculating unit calculates a line length of the epithelial boundary detected by the epithelial boundary detecting unit, and calculates the feature amount based on the calculated line length.
  The image processing apparatus for esophageal mucosa according to claim 1.   The feature amount calculating means calculates a sum value in the circumferential direction of an absolute value of an angle formed by adjacent vectors in the epithelial boundary curve detected by the epithelial boundary detecting means, and based on the calculated sum value, the feature Calculate quantity
  The image processing apparatus for esophageal mucosa according to claim 1.   The feature amount calculating means calculates the number of inflection points at the epithelial boundary detected by the epithelial boundary detection means, and calculates the feature amount based on the calculated number of inflection points.
  The image processing apparatus for esophageal mucosa according to claim 1.   The feature amount calculation means calculates the number of extreme values that are maximum or minimum at the epithelial boundary detected by the epithelial boundary detection means, and calculates the feature amount based on the calculated number of extreme values.
  The image processing apparatus for esophageal mucosa according to claim 1.   Development view generating means for generating a development view for an image of the inside of the esophagus;
  Epithelial boundary detection means for detecting an epithelial boundary as a boundary between the squamous epithelium as the esophageal mucosa and the columnar epithelium of the gastric mucosa based on the color tone change of the pixel value on the developed view or the edge state of the image ,
  A feature amount calculating means for calculating a feature amount relating to a coordinate with respect to a lumen direction of each pixel constituting the epithelial boundary detected by the epithelial boundary detecting means;
  Analysis means for outputting an analysis result based on the feature quantity calculated by the feature quantity calculation means;
  An image processing apparatus for esophageal mucosa, comprising:   The feature amount calculating means calculates the feature amount relating to a difference value between a maximum value and a minimum value in the lumen direction of the esophagus of pixels constituting the epithelial boundary.
  The image processing apparatus for esophageal mucosa according to claim 8.   Development view generating means for generating a development view for an image of the inside of the esophagus;
  Epithelial boundary detection means for detecting an epithelial boundary as a boundary between the squamous epithelium as the esophageal mucosa and the columnar epithelium of the gastric mucosa based on the color tone change of the pixel value on the developed view or the edge state of the image ,
  A reference part setting means for setting a reference part on the development view;
  A feature amount calculating means for calculating a feature amount relating to a distance between the coordinates of the pixels constituting the epithelial boundary detected by the epithelial boundary detecting means and the pixels constituting the reference portion set by the reference portion setting means; ,
  Analysis means for outputting an analysis result based on the feature quantity calculated by the feature quantity calculation means;
  An image processing device for esophageal mucosa, comprising:   The reference portion is a gastroesophageal junction in the esophagus;
  The feature amount calculating means calculates a feature amount relating to a distance between coordinates of a pixel constituting the epithelial boundary and a pixel constituting the gastroesophageal junction.
  The image processing apparatus for esophageal mucosa according to claim 10.

Images