JP6855610B2 - Image processing device, operation method of image processing device, and image processing program - Google Patents

Description

The present invention relates to an image processing device for processing an image or the like used at the time of diagnosing atrophic gastritis, an operation method of the image processing device, and an image processing program.

In the medical field, diagnosis using an endoscope system including a light source device, an electronic endoscope, and a processor device has become widespread. In particular, in hospitals that perform gastrointestinal endoscopy, a definitive diagnosis is required before performing endoscopic treatment such as endoscopic mucosal resection or endoscopic submucosal dissection, and the state of mucosal changes is observed. There is an increasing need to observe endoscopic images that emphasize changes in order to do so.

Observation methods that emphasize changes include optical methods, digital methods, optical digital methods, and dye spraying methods. FICE (Flexible spectral Imaging Color), which is a digital method, is used for observing and picking up microvessels in the surface structure. Enhancement (registered trademark), NBI (Narrow Band imaging, registered trademark), which is an optical digital method, or indigocamine spraying or rugor spraying, which is a dye spraying method, is used. Endoscopes generally take pictures using white light, but FICE extracts spectroscopic images by performing signal processing to make it easier to see the properties of tissues or blood vessels. On the other hand, NBI acquires and highlights the reflection spectrum information of a living body by narrowing the band of white light and irradiating it with an optical filter.

Originally, an endoscopic image tends to be concentrated in a red part and a narrow range of hues, and there is little difference between a lesion part and a normal part, and a process for emphasizing the difference is desired. Therefore, a hue-enhancing type that emphasizes hue and saturation, or a hue-enhancing type that emphasizes differences in hue has been proposed (for example, Patent Document 1). In Patent Document 1, in the hue-enhanced type, the color vector of the average color is obtained in a two-dimensional (RY and BY) color space excluding the influence of brightness from the image, and the color vector is radially centered on the position of the color vector. The range of hue is expanded to the moved position. On the other hand, in the hue-coordinated type, the hue is expanded around the position of the color vector. Further, it is disclosed that since the hue is different between the case where the dye is sprayed and the case where the dye is not sprayed, the table for conversion is changed between the case where the dye is sprayed and the case where the dye is not sprayed. There is.

In addition, in an endoscope that uses narrowed-band light, the endoscopy emphasizes the capillaries and fine patterns of the mucosa on the surface of the mucosa because it is observed by irradiating light with a wavelength that is easily absorbed by hemoglobin in the blood. It becomes an image. When the gastric mucosa in the endoscopic image is normal, the surface mucosal layer is thick and most of the light is absorbed and then reflected by this mucosal layer, which is within the normal submucosal gastric mucosa. The blood vessels can hardly be observed on the endoscopic image. On the other hand, in the case of gastric mucosa with advanced atrophic gastritis, the mucosal layer is thinned due to the decrease in gastric gland cells. On the image, the muscularis mucosae, which has a color close to white, can be seen through, and the color of the atrophic mucosa is faded from that of the normal part. Further, in the atrophic mucosal part, as the mucosal layer becomes thinner with atrophy, the blood vessels in the submucosal layer become transparent. Therefore, in the diagnosis of a gastric lesion based on atrophic gastritis, the degree of progression of atrophy and the boundary between the normal part and the gastritis part are determined by utilizing the above characteristics.

Japanese Unexamined Patent Publication No. 1-113018

However, even with an endoscope that uses narrowed-band light, if the atrophy is highly advanced (for example, in the case of atrophy included in groups C and D in ABC screening), the endoscopic image The above features can be clearly observed above, but if the atrophy is not very advanced (eg, in the case of atrophy contained in groups B and C on ABC screening), on an endoscopic image. The difference between the atrophied part and the normal part is small, and it may be difficult to judge the progress of the atrophy and the boundary between the normal part and the gastritis part. Therefore, it is required to clarify the difference due to the above features on the endoscopic image so that the boundary between the normal part and the gastritis part can be observed.

An object of the present invention is to provide an image processing device, an operation method of the image processing device, and an image processing program capable of emphasizing a color change of a mucous membrane or the like that may occur when the stomach is atrophied due to atrophic gastritis.

The image processing apparatus of the present invention has an image acquisition unit that acquires an endoscopic image composed of an image signal including a narrow band image signal, and a three-dimensional color including a coordinate axis of at least one component of brightness and brightness. A color difference vector that acquires a color difference vector, which is a three-dimensional vector in a three-dimensional color space, from a reference color representing the color of a normal part of an endoscopic image to the color of each pixel of interest in the endoscopic image in space. In the acquisition unit and this color space, the color of each attention pixel is added to the reference color by adding the color difference expansion vector of the vector amount which is in the same direction as the color difference vector of the attention pixel and which is the vector amount of the color difference vector plus the emphasis amount. A color difference expansion unit that performs color difference expansion processing to convert to color, and an output unit that outputs an endoscopic image that expands the color difference between the normal part and the abnormal part by applying the color difference expansion processing to all pixels of the endoscopic image. The color difference vector acquisition unit uses a reference color having a higher brightness value as the representative brightness value becomes higher, corresponding to a representative brightness value representing the endoscopic image among a plurality of predetermined reference colors. The lower the representative brightness value, the lower the brightness value of the reference color, and the color difference vector is acquired.

The method of operating the image processing device of the present invention is a method of operating an image processing device including an image acquisition unit, a color difference vector acquisition unit, a color difference expansion unit, and an output unit, and the image acquisition unit has a narrow band. The image acquisition step of acquiring the endoscopic image generated by the image signal including the image signal and the color difference vector acquisition unit are narrow in the three-dimensional color space including the coordinate axes of at least one component of brightness and brightness. A three-dimensional vector in a three-dimensional color space, from the reference color representing the color of the normal part of the endoscopic image generated by the image signal including the band image signal to the color of each noteworthy pixel in the endoscopic image. In this color space, the color difference vector acquisition step for acquiring a certain color difference vector and the color difference extension unit add an emphasis amount to the color difference vector of the color difference vector in the same direction as the color difference vector of the color difference vector of the attention pixel. The color difference expansion step that converts the color difference expansion vector of the vector amount to the color added to the reference color, and the normal part and abnormality that the output unit applied the color difference expansion processing to all the pixels of the endoscopic image. It includes an output step that outputs an endoscopic image with the color difference of the part expanded, and the color difference vector acquisition step corresponds to a representative brightness value representing the endoscopic image among a plurality of predetermined reference colors. A color difference vector is acquired by using a reference color having a higher brightness value as the representative brightness value becomes higher and using a reference color having a lower brightness value as the representative brightness value becomes lower.

The image processing program of the present invention includes a computer, an image acquisition unit that acquires an endoscopic image generated by an image signal including a narrow band image signal, and a coordinate axis of at least one component of brightness and brightness. Acquires a color difference vector, which is a three-dimensional vector in a three-dimensional color space, from a reference color representing the color of a normal part of an endoscopic image to the color of each pixel of interest in the endoscopic image in a dimensional color space. The color difference vector acquisition unit and the color difference expansion vector of the vector amount in which the color of each attention pixel is in the same direction as the color difference vector of the attention pixel and the emphasis amount is added to the vector amount of the color difference vector as the reference color in this color space. Output that outputs an endoscopic image that expands the color difference between the normal part and the abnormal part by applying the color difference expansion processing to all the pixels of the endoscopic image and the color difference expansion part that performs the color difference expansion processing to convert to the color added to The color difference vector acquisition unit functions as a unit, and the color difference vector acquisition unit selects a reference color having a higher brightness value as the representative brightness value becomes higher, corresponding to a representative brightness value representing an endoscopic image among a plurality of predetermined reference colors. The color difference vector is acquired by using the reference color whose brightness value is lower as the representative brightness value becomes lower.

The "narrow band image signal" is an image signal obtained by imaging a sample irradiated with narrow band light, and the "image signal including a narrow band image signal" is a sample irradiated with light containing narrow band light. Refers to an image signal obtained by imaging.

The "color space" refers to a space in which colors can be indicated by coordinates, and the "color space including brightness" refers to a three-dimensional color space including a component of brightness. Luminance refers to a color component that represents the degree of brightness. Like brightness, lightness is also used as a component indicating the degree of brightness, and the color space includes a color space having coordinate axes of components represented by brightness and a color space having coordinate axes of components represented by lightness. However, in the present specification, neither brightness nor lightness is distinguished as a component indicating the degree of brightness.

The "color difference vector" refers to a three-dimensional vector representing the difference between a vector representing the position of a reference color in a three-dimensional color space and a vector representing the position of the color of a pixel of interest.

Further, when the vector amount of the color difference vector is smaller than the first threshold value, the color difference extension unit increases the emphasis amount as the vector amount of the color difference vector becomes larger, and when the vector amount of the color difference vector is larger than the first threshold value, the emphasis amount is increased. , It is preferable that the emphasis amount becomes smaller as the vector amount becomes larger.

Further, a storage unit that stores a plurality of reference colors in advance is further provided, and the color difference vector acquisition unit is represented by a representative luminance value representing an endoscopic image among the plurality of reference colors of the storage unit. The color difference vector may be acquired by using a reference color having a higher luminance value as the luminance value becomes higher and using a reference color having a lower luminance value as the representative luminance value becomes lower.

Further, the representative luminance value of the endoscopic image may be the average value of the luminance values of all the pixels of the endoscopic image.

Further, the reference color may be an average value of all the pixels of the endoscopic image.

Further, it is desirable that the reference color is calculated from the pixels of the endoscopic image other than the pixels having at least one color component exceeding the second threshold value among the color components of each element of the endoscopic image.

The "pixel having at least one or more color components exceeding the second threshold value" is a color component exceeding the color of the image that normally appears when the sample is photographed, and is, for example, the color of a portion causing halation or the like. Pixels that have components or pixels that have color components that appear when something other than a sample is photographed.

Further, the color space may be any of Ycc color space, Lab color space, Luv color space, HSB color space, HSV color space, and HSV color space.

Further, it is desirable that the narrow band image signal is obtained by imaging a sample illuminated with narrow band light that absorbs more light into blood than other bands.

Further, the narrow band image signal is a blue narrow band image signal obtained by imaging a sample illuminated with blue narrow band light that absorbs more light into blood than other bands in the blue band, or in the green band. It may be a green narrow band image signal obtained by imaging a sample illuminated with green narrow band light that absorbs more light into blood than other bands.

Further, when the endoscopic image is an image of the inner wall of the stomach, the normal part is the normal mucosa and the abnormal part is the abnormal mucosa.

Further, the other image processing apparatus of the present invention includes an image acquisition unit that acquires an endoscopic image composed of an image signal including a narrow band image signal, and a plurality of reference colors that represent colors of a predetermined normal portion. Of these, the standard color with a higher brightness value is used as the average value is higher, and the standard color with a lower brightness value is used as the average value is lower, corresponding to the average value of the brightness values of the endoscopic image. In the three-dimensional color space including the coordinate axes of at least one of the components, the color difference vector, which is a three-dimensional vector in the three-dimensional color space from the reference color to the color of each attention pixel in the endoscopic image, is acquired. In the color space and the color difference vector acquisition unit, the color of each attention pixel is set in the same direction as the color difference vector of the attention pixel, and the color difference expansion vector of the vector amount obtained by adding the emphasis amount to the vector amount of the color difference vector is used as the reference color. A color difference expansion unit that performs color difference expansion processing to convert to the added color, and an output unit that outputs an endoscopic image that expands the color difference between the normal part and the abnormal part by applying the color difference expansion processing to all pixels of the endoscopic image. The color difference expansion unit has the distance between the color of the abnormal portion in the color space and the reference color as the first threshold value, and when the vector amount of the color difference vector is smaller than the first threshold value, the vector amount of the color difference vector is The larger the amount, the larger the emphasis amount, and when the vector amount of the color difference vector is larger than the first threshold value, the larger the vector amount, the smaller the emphasis amount.

Further, another method of operating the image processing device of the present invention is a method of operating an image processing device including an image acquisition unit, a color difference vector acquisition unit, a color difference expansion unit, and an output unit, and is an image acquisition unit. However, the image acquisition step of acquiring the endoscopic image generated by the image signal including the narrow band image signal, and the color difference vector acquisition unit are used to perform endoscopy among a plurality of reference colors representing the colors of the predetermined normal parts. Corresponding to the average value of the brightness values of the mirror image, the higher the average value, the higher the brightness value of the reference color, and the lower the average value, the lower the brightness value of the reference color, at least one of the brightness and the brightness. Color difference vector acquisition to acquire the color difference vector, which is a three-dimensional vector in the three-dimensional color space from the reference color to the color of each noteworthy pixel in the endoscopic image in the three-dimensional color space including the coordinate axes of the two components. In the color space, the step and the color difference expansion unit set the color of each attention pixel in the same direction as the color difference vector of the attention pixel, and the color difference expansion vector of the vector amount obtained by adding the emphasis amount to the vector amount of the color difference vector as the reference color. A color difference expansion step that performs color difference expansion processing to convert to the color added to, and an endoscopic image in which the output unit applies color difference expansion processing to all pixels of the endoscopic image to expand the color difference between the normal part and the abnormal part. In the color difference expansion step, the distance between the color of the abnormal part in the color space and the reference color is set as the first threshold, and when the vector amount of the color difference vector is smaller than the first threshold, the color difference is provided. The larger the vector amount of the vector, the larger the emphasis amount, and when the vector amount of the color difference vector is larger than the first threshold value, the larger the vector amount, the smaller the emphasis amount.

In addition, other image processing programs of the present invention include an image acquisition unit that acquires an endoscopic image composed of an image signal including a narrow band image signal, and a plurality of image processing programs that represent a predetermined color of a normal part. Corresponding to the average value of the brightness values of the endoscopic image, the reference color having a higher brightness value is used as the average value is higher, and the reference color having a lower brightness value is used as the average value is lower. And in a three-dimensional color space that includes the coordinate axes of at least one component of lightness, a color difference that is a three-dimensional vector in the three-dimensional color space from the reference color to the color of each noteworthy pixel in the endoscopic image. A color difference vector acquisition unit that acquires a vector, and a color difference expansion vector of a vector amount in which the color of each attention pixel in the color space is in the same direction as the color difference vector of the attention pixel and the emphasis amount is added to the vector amount of the color difference vector. Outputs a color difference expansion part that performs color difference expansion processing to convert to a color added to the reference color, and an endoscope image that expands the color difference between the normal part and abnormal part where the color difference expansion processing is applied to all pixels of the endoscopic image. An image processing program for functioning as an output unit, in which the color difference expansion unit sets the distance between the color of the abnormal portion in the color space and the reference color as the first threshold, and the vector amount of the color difference vector is the first threshold. If it is smaller, the emphasis amount is increased as the vector amount of the color difference vector is larger, and if the vector amount of the color difference vector is larger than the first threshold value, the emphasis amount is decreased as the vector amount is larger.
Further, the other image processing apparatus of the present invention includes an image acquisition unit that acquires an endoscopic image composed of an image signal including a narrow band image signal, and a coordinate axis of at least one component of brightness and brightness. In the three-dimensional color space, the color difference vector, which is a three-dimensional vector in the three-dimensional color space, from the reference color representing the color of the normal part of the endoscopic image to the color of each attention pixel in the endoscopic image. The color difference vector acquisition unit to be acquired and the color difference expansion vector of the vector amount in which the color of each attention pixel in the color space is in the same direction as the color difference vector of the attention pixel and the emphasis amount is added to the vector amount of the color difference vector. Outputs a color difference expansion part that performs color difference expansion processing to convert to a color added to the reference color, and an endoscope image that expands the color difference between the normal part and abnormal part where the color difference expansion processing is applied to all pixels of the endoscopic image. The reference color is the average value of the pixels of the endoscopic image other than the pixels having at least one or more color components exceeding the second threshold value among the color components of each pixel of the endoscopic image. And.

The "color of the abnormal part" means a color that is likely to appear in the abnormal part, and means a color that represents the abnormal part. For example, it may be the color that appears most frequently in the abnormal portion, or it may be the color obtained by averaging the colors of the abnormal portion.

The "distance between the color of the abnormal portion and the reference color in the color space" refers to the vector amount of the difference vector between the reference color and the color of the abnormal portion obtained in the color space.

According to the present invention, the color of each attention pixel of the endoscopic image composed of the image signal including the narrow band image signal is the reference color representing the color of the normal portion and each attention pixel in the color space including the brightness. By emphasizing the color difference in the direction of the color difference vector that represents the color difference, for example, the color difference between the abnormal part such as the mucous membrane and the normal part that may occur when the stomach is atrophic due to atrophic gastric inflammation is taken into consideration, and the difference in brightness is also taken into consideration. Can be emphasized.

External view of the endoscopic system. The block diagram which shows the function of the endoscope system of 1st Embodiment. The graph which shows the spectral intensity of white light. A graph showing the spectral intensity of special light. The block diagram which shows the internal structure of the abnormality area emphasis part. The figure for demonstrating the color difference enhancement in the color space which has the axis of the luminance component. The figure which shows the relationship between the vector amount and the emphasis amount of a color difference vector. The figure for demonstrating the difference between the color difference enhancement of an abnormal part and the color difference emphasis of a bleeding part in a color space. The figure for demonstrating the difference between the color difference enhancement of an abnormal part in a color space and the color difference emphasis of a color close to a normal part. A flowchart showing the flow of diagnosis of atrophic gastritis. The block diagram which shows the function of the endoscope system of 2nd Embodiment. A graph showing the wavelength of a light source.

As shown in FIG. 1, the endoscope system 10 of the first embodiment includes an endoscope 12, a universal code 13, a light source device 14, a processor device 16, a monitor 18, and an input device 20. Have. The endoscope 12 is optically connected to the light source device 14 and electrically connected to the processor device 16 via the universal cord 13. The endoscope 12 has an insertion portion 21 to be inserted into the sample, an operation portion 22 provided at the base end portion of the insertion portion, and a curved portion 23 and a tip portion 24 provided on the tip end side of the insertion portion 21. doing. By operating the angle knob 22a of the operation unit 22, the bending unit 23 bends. Along with this bending motion, the tip portion 24 is directed in a desired direction.

In addition to the angle knob 22a, the operation unit 22 is provided with a mode changeover switch (mode changeover SW) 22b and a zoom operation part 22c. The mode changeover switch 22b is used for a changeover operation between two types of modes, a normal observation mode and a special observation mode. The normal observation mode is a mode in which white light is used to illuminate the inside of the sample. The special observation mode is a mode in which a bluish special light is used to illuminate the inside of the sample, and is a mode in which the change in the color of the mucous membrane and the see-through of blood vessels that may occur when the stomach is atrophied due to atrophic gastritis are emphasized. The zoom operation unit 22c is used for a zoom operation for enlarging a sample by driving a zooming lens 47 (see FIG. 2) in the endoscope 12.

The processor device 16 is connected to the monitor 18 and the input device 20. The monitor 18 displays image information and the like. The input device 20 functions as a user interface that accepts input operations such as function settings. An external recording unit (not shown) for recording image information or the like may be connected to the processor device 16.

As shown in FIG. 2, the light source device 14 includes a blue laser light source (445LD) 34 that emits a blue laser light having a central wavelength of 445 nm and a blue-violet laser light source (405LD) 36 that emits a blue-violet laser light having a central wavelength of 405 nm. It is provided as a light source. The light emitted from the semiconductor light emitting elements of each of the light sources 34 and 36 is individually controlled by the light source control unit 40, and the light amount ratio between the emitted light of the blue laser light source 34 and the emitted light of the blue-violet laser light source 36 can be freely changed. It has become. In the normal observation mode, the light source control unit 40 mainly drives the blue laser light source 34 and controls so that the blue-violet laser light is slightly emitted. In the case of this normal observation mode, the blue-violet laser light source 36 may be driven. However, in this case, it is preferable to keep the emission intensity of the blue-violet laser light source 36 low.

On the other hand, in the special observation mode, both the blue laser light source 34 and the blue-violet laser light source 36 are driven, and the emission ratio of the blue laser light is made larger than the emission ratio of the blue-violet laser light. I'm in control. The half-value width of the blue laser light or the blue-violet laser light is preferably about ± 10 nm. Further, as the blue laser light source 34 and the blue-violet laser light source 36, a broad area type InGaN-based laser diode can be used, and an InGaAsN-based laser diode or a GaAsN-based laser diode can also be used. Further, as the light source, a light emitting body such as a light emitting diode may be used.

The laser light emitted from each of the light sources 34 and 36 is incident on the light guide (LG) 41 via an optical member (none of which is shown) such as a condenser lens, an optical fiber, and a combiner. The light guide 41 is built in the light source device 14, the endoscope 12, and the universal cord (cord for connecting the endoscope 12 and the light source device 14) 13. The blue laser light having a central wavelength of 445 nm or the blue-violet laser light having a central wavelength of 405 nm is propagated to the tip portion 24 of the endoscope 12 via the light guide 41. A multimode fiber can be used as the light guide 41. As an example, a fine fiber cable having a core diameter of 105 μm, a clad diameter of 125 μm, and a diameter of φ0.3 to 0.5 mm including a protective layer serving as an outer skin can be used.

Such blue narrow-band light from the blue laser light and the blue-violet laser light is highly absorbed by the light-absorbing substance in the mucous membrane, specifically, blood (particularly hemoglobin) which is abundant in the digestive organs, and is in the special observation mode. The difference between the normal mucosal region and the atrophic mucosal region becomes large when the image is taken.

The tip 24 of the endoscope 12 has an illumination optical system 24a and an imaging optical system 24b. The illumination optical system 24a is provided with a phosphor 44 into which a blue laser light having a center wavelength of 445 nm or a blue-violet laser light having a center wavelength of 405 nm is incident from the light guide 41, and an illumination lens 45. When the phosphor 44 is irradiated with a blue laser beam, fluorescence is emitted from the phosphor 44. Further, some blue laser light passes through the phosphor 44 as it is. The bluish-purple laser light passes through the phosphor 44 without exciting it. The light emitted from the phosphor 44 is irradiated into the sample through the illumination lens 45.

Here, in the normal observation mode, since the blue laser light is mainly incident on the phosphor 44, the blue laser light as shown in FIG. 3A and the fluorescence excited and emitted from the phosphor 44 by the blue laser light are combined. White light is emitted into the sample. On the other hand, in the special observation mode, since both the blue-violet laser light and the blue laser light are incident on the phosphor 44, the phosphor is generated by the blue-violet laser light, the blue laser light, and the blue laser light as shown in FIG. 3B. The sample is irradiated with special light that is a combination of fluorescence that is excited and emitted from 44. In this special observation mode, since the blue component contains a blue-violet laser light in addition to the blue laser light having a high emission intensity, the special light contains a large amount of the blue component and the wavelength range covers almost the entire visible light range. It is wideband light.

The phosphor 44 is a plurality of types of phosphors (for example, YAG (yttrium aluminum garnet) -based phosphors, BAM (BaMgAl10O17), etc.) that absorb a part of blue laser light and excite and emit light from green to yellow. It is preferable to use one containing (fluorescent material of). When a semiconductor light emitting element is used as an excitation light source for the phosphor 44 as in this configuration example, high-intensity white light can be obtained with high luminous efficiency, the intensity of white light can be easily adjusted, and the color of white light can be adjusted. Changes in temperature and chromaticity can be suppressed to a small extent.

As shown in FIG. 2, the image pickup optical system 24b of the endoscope 12 includes an image pickup lens 46, a zooming lens 47, and an image pickup sensor 48. The reflected light from the sample enters the image sensor 48 via the image pickup lens 46 and the zooming lens 47. As a result, a reflected image of the sample is formed on the image sensor 48. The zooming lens 47 moves between the tele end and the wide end by operating the zoom operation unit 22c. When the zooming lens 47 moves to the wide end side, the reflected image of the sample is reduced, while when it is moved to the tele end side, the reflected image of the sample is enlarged.

The image sensor 48 is a color image sensor that captures a reflected image of a sample and outputs an image signal. The image sensor 48 is preferably a CCD (Charge Coupled Device) image sensor, a CMOS (Complementary Metal-Oxide Semiconductor) image sensor, or the like. The image sensor used in the present invention is an RGB image sensor having RGB channels provided with an RGB color filter on the imaging surface, and an R (red) color filter is provided by performing photoelectric conversion on each channel. The R image signal is output from the pixels, the G image signal is output from the G pixel provided with the G (green) color filter, and the B image signal is output from the B pixel provided with the B (blue) color filter. ..

The image sensor 48 may be an image sensor having C (cyan), M (magenta), Y (yellow), and G (green) CMYG filters on the image pickup surface. In the case of an image sensor provided with a CMYG filter, it is possible to obtain an image signal of three colors of RGB by color conversion from the image signal of four colors of CMYG. In this case, it is necessary that either the endoscope 12, the light source device 14, or the processor device 16 is provided with a color conversion means for color-converting a CMYG four-color image signal into an RGB three-color image signal. is there.

The image signal output from the image sensor 48 is transmitted to the CDS (correlated double sampling) / AGC (Automatic Gain Control) circuit 50. The CDS / AGC circuit 50 performs correlated double sampling (CDS) and automatic gain control (AGC) on an image signal which is an analog signal. The image signal that has passed through the CDS / AGC circuit 50 is gamma-converted by the gamma conversion unit 51. As a result, an image signal having a gradation suitable for an output device such as a monitor 18 can be obtained. The image signal after the gamma conversion is converted into a digital image signal by the A / D converter (A / D converter) 52. The A / D converted digital image signal is input to the processor device 16.

The processor device 16 includes a receiving unit 54, an image processing switching unit 60, a normal optical image processing unit 62, a special optical image processing unit 64, and an image display signal generation unit 66. The receiving unit 54 receives the digital image signal from the endoscope 12. The receiving unit 54 includes a DSP (Digital Signal Processor) 56 and a noise removing unit 58. The DSP 56 performs gamma correction and color correction processing on the digital image signal. The noise removing unit 58 removes noise from the digital image signal by performing noise removing processing (for example, a moving average method, a median filter method, etc.) on the digital image signal that has been gamma-corrected by the DSP 56. The digital image signal from which noise has been removed is transmitted to the image processing switching unit 60.

The image processing switching unit 60 transmits a digital image signal to the normal optical image processing unit 62 when the mode changeover switch 22b is set to the normal observation mode, and when the mode changeover switch 22b is set to the special observation mode, the image processing switching unit 60 transmits a digital image signal. The digital image signal is transmitted to the special optical image processing unit 64.

The normal optical image processing unit 62 includes an image acquisition unit 68, a color enhancement unit 70, and a structure enhancement unit 72. The image acquisition unit 68 allocates the input RGB3 channel digital image signal to the R image data, the G image data, and the B image data, respectively, and acquires the RGB image data. Further, the RGB image data may be subjected to color conversion processing such as 3 × 3 matrix processing, gradation conversion processing, and three-dimensional LUT processing.

The color enhancement unit 70 performs various color enhancement processes on the RGB image data. The structure enhancement unit 72 performs structure enhancement processing such as spatial frequency enhancement on the RGB image data. The RGB image data subjected to the structure enhancement processing by the structure enhancement unit 72 is input from the normal optical image processing unit 62 to the image display signal generation unit 66 as a normal light image.

The special optical image processing unit 64 includes an image acquisition unit 74, an abnormal region enhancement unit 77, and a structure enhancement unit 78. The image acquisition unit 74 acquires RGB image data in the same manner as the image acquisition unit 68. The abnormal region emphasizing portion 77 performs a color difference expanding process for emphasizing the color difference between the normal mucosal region (normal portion) and the abnormal region (abnormal portion) that may include a lesion portion such as atrophic mucosa or gastric cancer. The details of the abnormal region emphasizing unit 77 will be described later. The structure enhancement unit 78 performs structure enhancement processing such as spatial frequency enhancement on the RGB image data that has undergone the color difference expansion processing. The RGB image data subjected to the structure enhancement processing by the structure enhancement unit 78 is input from the special light image processing unit 64 to the image display signal generation unit 66 as a special light image.

The image display signal generation unit 66 converts a normal light image or a special light image input from the normal light image processing unit 62 or the special light image processing unit 64 into a display image signal for displaying as a displayable image on the monitor 18. To do. The converted display image signal is output to the monitor 18, and the monitor 18 displays the display image signal as a normal light image or a special light image.

As shown in FIG. 4, the abnormal region emphasizing unit 77 includes a color space conversion unit 80, a color difference vector acquisition unit 82, a color difference expansion unit 84, a color space inverse conversion unit 86, and a storage unit 88. When the gastric mucosa on the inner wall of the stomach in the endoscopic image is a normal part, the blood vessels in the normal submucosa of the gastric mucosa can hardly be observed on the endoscopic image, but abnormal mucosa such as atrophic mucosa. In the case of (abnormal part), the mucosal layer becomes thin due to the decrease of gastric gland cells, and the mucosal muscle plate of a color close to white can be seen through, so that the color of the abnormal part is faded from that of the normal part. However, endoscopic images of the stomach wall tend to be concentrated in the red part and in a narrow area in the color space, and the difference in color between the normal part and the abnormal part is not seen by a skilled observer. It is difficult to distinguish visually. Therefore, using the color of the normal mucous membrane as a reference color, a color difference expansion process for expanding the color difference according to the difference between the color of each pixel of the image data and the reference color is performed.

First, the color space conversion unit 80 converts the R value, G value, and B value of each pixel represented in the RGB color space of the RGB image data acquired by the image acquisition unit 74 into three dimensions having a brightness or brightness component. Convert to a value in the color space of. The color space having a brightness or lightness component includes a Y-axis representing a brightness component, a Ycc color space having a Cb axis and a Cr axis representing a color difference component, and an L-axis representing a lightness component and an a-axis representing a complementary color component. Lab color space with b-axis, L-axis representing lightness component, Luv color space with u-axis and v-axis representing saturation and hue components, hue H / saturation S / lightness B (or lightness V) There is an HSB color space (or HSV color space) consisting of the axes of the components of, or an HSL color space consisting of the axes of the components of hue H, saturation S, and brightness L, but in a color space containing the components of brightness or lightness. Any color space may be used as long as it is available. Further, it may be an RGB color space having no axis representing the luminance or brightness component. Hereinafter, in the present embodiment, a case where the color difference expansion processing is performed after converting to a color space having an axis representing the luminance or lightness component will be described. Further, although some of the color spaces after conversion have axes representing the components of luminance and brightness, they will be described below as luminance for convenience without distinguishing between luminance and brightness.

The color difference vector acquisition unit 82 determines the color difference from the reference color representing the color of the normal mucous membrane, which is the normal part of the endoscopic image, to the color of each attention pixel in the endoscopic image, in a color space having a brightness axis. Then, it is acquired as a three-dimensional color difference vector obtained by subtracting the vector of the position of the reference color from the vector of the position of the color of each pixel of interest.

The color of the normal mucosa differs depending on whether the entire endoscopic image is a bright image or a dark image. Therefore, as the reference color, a plurality of reference colors corresponding to the brightness value of the endoscopic image are stored in advance in the storage unit 88, and the reference color corresponding to the brightness of the endoscopic image is used as the reference color and each of them. Get the color difference vector of the pixel color. For these reference colors, colors that are likely to statistically appear on the normal mucosa of the endoscopic image from the endoscopic images taken in the past are determined according to various brightness values of the endoscopic image. Can be done. Specifically, when the brightness value of the endoscopic image is high, the brightness value of the reference color is also high, and when the brightness value of the endoscope image is low, the brightness value of the reference color is also low. Therefore, a table in which different reference colors are set corresponding to a plurality of luminance values is stored in the storage unit 88, and the reference color corresponding to the luminance value closest to the endoscopic image to be subjected to the color difference expansion processing is stored. To take out. Alternatively, when the reference color when the brightness value is high and the reference color when the brightness value is low are stored and the brightness value of the endoscopic image that is the target of the color difference expansion processing is the brightness value in between. Is obtained by interpolating the reference color when the brightness value is high and the reference color when the brightness value is low, and when the brightness value of the endoscopic image is out of the range, the brightness value is The reference color when the brightness value is high and the reference color when the brightness value is low may be obtained by extrapolation interpolation.

Further, as the brightness value of the endoscope image, a representative brightness value representing the brightness of all the pixels of the endoscope image can be used. For example, the average value of the brightness values of all the pixels of the endoscopic image or the median value of the brightness values of all the pixels of the endoscopic image may be used as the representative brightness value.

In order to clarify the difference between the normal part which is the normal mucous membrane and the abnormal part which is the abnormal mucous membrane, the color difference expansion unit 84 sets the color of each attention pixel in the color space having the axis of brightness by the color difference vector acquisition unit 82. The color difference extension vector of the vector amount obtained by adding the enhancement amount to the vector amount of the color difference vector in the same direction as the acquired color difference vector of each attention pixel is converted into the color difference enhancement color obtained by adding the reference color to the normal part. Increase the color difference of the abnormal part.

With reference to FIG. 5, the color difference enhancement in the Lab color space having the axis of the luminance (brightness) component will be specifically described. In FIG. 5, the L-axis is the brightness (brightness) component, and the a-axis and the b-axis are complementary color components. When the position S of the reference color in the Lab color space and the position of the color of the pixel of interest are P, the color difference vector is p. As shown in the following equation (1), the emphasis amount d corresponding to the vector amount | p | of the color difference vector p is changed to the color difference enhancement color Q at the position extended in the direction of the color difference vector p. In this way, in the Lab color space, the difference between the abnormal part and the normal part is obtained by expanding the difference between the reference color and the color of each pixel in the same direction as the color difference vector representing the difference between the reference color and the color of each pixel. Can be clarified naturally.

Further, since it is preferable to maintain the same color for pixels having a color close to the standard color, that is, pixels having a color close to the normal mucosa, the emphasis amount d is reduced, and the part close to the color of the abnormal mucosa is regarded as the normal mucosa. It is preferable to increase the emphasis amount d so that the difference between the two is clear. On the other hand, it is preferable not to emphasize the part that greatly deviates from the color of the mucous membrane because it is highly likely that the color is not the mucosal color such as the bleeding part or the part where the pigment is sprayed.

FIG. 6 shows an example showing the relationship between the emphasis amount d and the vector amount | p |. As shown in FIG. 5, the color difference vector between the typical color and the reference color appearing on the abnormal mucous membrane so that the color changes naturally as a whole and the difference between the color appearing on the abnormal mucous membrane and the reference color becomes clear. A value close to the vector amount (distance) | p | of p is set as the first threshold value R1, and when the vector amount | p | of the color difference vector p is smaller than the first threshold value R1, the vector amount of the color difference vector p | p | When the vector amount | p | of the color difference vector p is larger than the first threshold R1, the emphasis amount d is increased as the vector amount | p | is larger, and the emphasis amount d is decreased as the vector amount | p | is larger. Specifically, for each reference color, a lookup table in which the correspondence between the vector amount | p | of the color difference vector and the emphasis amount d is determined in advance is stored in the storage unit 88, and the reference color S and the color difference vector p are stored in advance. The emphasis amount d corresponding to the vector amount | p | of is drawn out, and the color subjected to the color difference expansion processing is calculated. By increasing the emphasis amount when the vector amount of the color difference vector is near the first threshold value R1, the color difference between the color of the abnormal mucosa and the color of the normal mucosa (reference color) is expanded to expand the color difference of the normal mucosa. And do not change the color that is clearly different from the normal mucosal color. This makes it easier to recognize the abnormal part, and the color of the normal part such as the normal mucosa and the part other than the mucous membrane where the image is taken can be observed as the original color.

Using FIGS. 7A and 7B, the color difference of each pixel after conversion when the emphasis amount is changed according to the vector amount of the color difference vector between the position S of the reference color in the Lab color space and the color position of each pixel. The position of the emphasized color in the color space will be described. P1 of FIG. 7A shows the position in the color space of the color of the portion considered to be the abnormal mucosa, and the emphasis amount d1 is increased and the color difference enhancement color is Q1 in the vicinity of the color close to the abnormal mucosa. On the other hand, P2 in FIG. 7A shows the position in the color space of the part considered to be the bleeding part due to the large color difference from the normal mucosa, and since it is a color different from the normal mucosal color, the emphasis amount d2 (<d1). ) Is reduced and the color difference enhancement color is Q2.

Similar to FIG. 7A, P1 in FIG. 7B indicates the position in the color space of the part considered to be the abnormal mucosa, and the emphasis amount d1 is increased in the vicinity of the color close to the abnormal mucosa, and the color difference enhancement color is Q1. To. On the other hand, P3 in FIG. 7B shows the position in the color space of the color close to the normal mucous membrane, and since the color difference from the reference color S is small and the color is close to the color of the normal part, the emphasis amount d3 (<d1) is reduced. The color difference emphasis color is set to Q3.

The color space inverse conversion unit 86 reconverts the color difference enhanced color obtained by the color difference expansion unit 84 into RGB image data. Further, if necessary, gamma conversion is performed on the RGB image data. As a result, RGB image data with color difference enhancement having gradation suitable for an output device such as a monitor 18 can be obtained. The output unit of the present invention is composed of a color space inverse conversion unit 86 and an image display signal generation unit 66.

Next, the operation and processing flow of the endoscope during a series of examinations in the present embodiment will be described with reference to the flowchart of FIG. First, the normal observation mode is set (S1), and the insertion portion 21 of the endoscope 12 is inserted into the sample. Normally, the tip 24 is advanced into the sample while observing an optical image (S2-N), and when the tip 24 of the insertion portion 21 reaches the stomach (S2-Y), it is diagnosed whether or not atrophic gastritis has occurred. (S3). Here, from the normal optical image, the mucosa is fading, or the boundary between the part where the dendritic deep blood vessel is transparent and the part where it is not transparent (called the endoscopic gland boundary). If it can be read (S3-Y), the doctor determines that it is a pathological finding that a lesion such as gastric cancer is caused by atrophic gastritis (S11).

On the other hand, if the presence of fading mucosa or endoscopic gland boundary cannot be read from the normal optical image (S3-N), the mode is switched to make a more reliable diagnosis. The switch 22b is operated to switch to the special observation mode (S4). By switching the special observation mode, special light including both blue laser light and blue-violet laser light is emitted. The image acquisition unit 74 acquires RGB image data from the RGB image signal obtained at the time of this special light emission.

First, the color space conversion unit 80 converts the R value, G value, and B value of each pixel represented in the RGB color space into, for example, a value in the Lab color space (S5). The representative luminance value is obtained from the luminance values (L values) of the pixels of all the endoscopic images. Next, the reference color corresponding to the representative luminance value is taken out from the storage unit 88, and the color difference vector between each pixel and the reference color is calculated (S6). Further, the lookup amount corresponding to the reference color is used to determine the emphasis amount according to the vector amount of the color difference vector. The color of each pixel is converted into a color difference enhancement color obtained by adding a color difference expansion vector of a vector amount obtained by adding an emphasis amount in the same direction as the color difference vector to a reference color (S7).

Based on the image data converted into the color difference emphasized color, the color space inverse conversion unit 86 returns the image data to the RGB color space, and the special light image is displayed on the monitor 18 (S8).

On the special light image, the mucosa is displayed in normal color when there is no atrophy of the stomach. In this case (S9-N), the doctor determines that there is no occurrence of a lesion such as gastric cancer due to atrophic gastritis (S10). On the other hand, when the atrophy of the stomach is progressing even slightly, the color of the atrophied mucosa is displayed in a fading tone (S9-Y). This makes it possible to clearly display the endoscopic gland boundary. Therefore, even if the color of the atrophic mucosa is not so fading in the actual stomach, the doctor determines that it is a pathological finding that a lesion such as gastric cancer is caused by atrophic gastritis. You will be able to do it (S11).

In the above description, the case where the reference color is determined according to the brightness value of the endoscopic image has been described, but the average value of the colors of all the pixels of the endoscopic image may be used. However, the reference color in this case deviates from the color component considered to be different from the mucosal color when at least one color component of all the pixel color components of the endoscopic image exceeds a specific value. The average value is calculated by using the second threshold value for the color having the value and excluding the pixels having the color component exceeding the second threshold value. The part where the color different from the mucous membrane color appears is, for example, the color that appears when a pixel with the color component of the part causing halation, the residue, the part where the pigment is sprayed, or the part other than the sample such as the bleeding part is photographed. It is a pixel having a component. The determination of the pixel having a component different from the mucosal color among the color components may be determined by the RGB value, or may be determined after converting to a color space having a luminance component. Alternatively, pixels having a color different from the color of the mucous membrane may be determined from the combination of RGB values, and the average value may be calculated by excluding the pixels having a color different from the color of the mucous membrane. Alternatively, a pixel having a color different from the color of the mucous membrane may be determined from the combination of the color components after conversion into the color space having the luminance component.

In the above embodiment, after converting the RGB color space into a color space having an axis representing a brightness or lightness component, a color difference expansion process is performed, and the color subjected to the color difference expansion process is converted into an RGB color space again. Although the case has been described, the color difference expansion process may be performed by the color difference vector acquisition unit 82 and the color difference expansion unit 84 in the RGB color space without converting from the RGB color space to another color space.

In the above embodiment, the case where special light including blue narrow band light (blue laser light and bluish purple laser light) having high light absorption with respect to the absorption substance of the mucous membrane is used has been described, but the absorbing substance of the mucous membrane has been described. On the other hand, light containing green narrow band light having high light absorption (for example, a wavelength component of 540 to 560 nm) may be used.

In the above-described embodiment, the case where the light source is a semiconductor light emitting device has been described, but in the second embodiment, an endoscope using an LED as the light source will be described with reference to FIGS. 9 and 10. In the second embodiment, the configurations of the light source device 14 of the endoscope and the illumination optical system 24c of the tip portion 24 are almost the same as those of the first embodiment, and therefore the same reference numerals are given and detailed description thereof will be omitted. To do. Further, since the operation and processing flow of the endoscope at the time of examination are almost the same as those of the first embodiment, detailed description thereof will be omitted.

As shown in FIG. 9, the light source device 14 includes a V-LED (Violet Light Emitting Diode) 42a, a B-LED (Blue Light Emitting Diode) 42b, a G-LED (Green Light Emitting Diode) 42c, and an R-LED (Red). The light source control unit 40 may control the drive of the Light Emitting Diode) 42d and the LEDs 42a to 42d of these four colors. In this configuration, an optical path coupling portion 43 for coupling the optical paths of the four colors of light emitted from the four-color LEDs 42a to 42d is provided, and the light coupled by the optical path coupling portion 43 is a light inserted into the insertion portion 21. The subject is irradiated through the guide (LG) 41 and the illumination lens 45. An LD (Laser Diode) may be used instead of the LED.

As shown in FIG. 10, the V-LED42a generates violet light Vi having a center wavelength of 405 ± 10 nm and a wavelength range of 380 to 420 nm. The B-LED42b generates blue light Bl with a center wavelength of 460 ± 10 nm and a wavelength range of 420 to 500 nm. The G-LED42c generates green light Gr having a wavelength range of 480 to 600 nm. The R-LED42d generates red light Re having a center wavelength of 620 to 630 nm and a wavelength range of 600 to 650 nm.

The light source control unit 40 lights the V-LED42a, B-LED42b, G-LED42c, and R-LED42d in both the normal observation mode and the special observation mode. Therefore, the observation target is irradiated with light in which four colors of purple light Vi, blue light Bl, green light Gr, and red light Re are mixed. Further, the light source control unit 40 sets each LED 42a to 42d so that the light amount ratio between the purple light Vi, the blue light Bl, the green light Gr, and the red light Re is Vic: Blc: Grc: Rec in the normal observation mode. Control. On the other hand, in the special observation mode, the light source control unit 40 sets each LED 42a to 42d so that the light amount ratio between the purple light Vi, the blue light Bl, the green light Ge, and the red light Re is Vis: Bls: Ges: Res. Control.

The tip 24 of the endoscope 12 is provided with an illumination optical system 24c and an imaging optical system 24b. The illumination optical system 24c has an illumination lens 45, and the light from the light guide 41 is irradiated to the observation target through the illumination lens 45. The imaging optical system 24b has substantially the same configuration as that of the first embodiment.

In the first and second embodiments described above, the case where the processing of the abnormal area emphasizing unit 77 is performed by the processor device 16 has been described, but the processing of the abnormal area emphasizing unit 77 in the processor device 16 is performed by a computer placed outside. An image processing program for functioning as the above-mentioned image acquisition unit, color difference vector acquisition unit, color difference expansion unit, and output unit is installed in the processor device 16 and recorded in an external recording unit or the like on an external computer. The processing of the abnormal region emphasizing unit 77 may be executed on the endoscopic image.

10 Endoscope system 12 Endoscope 13 Universal cord 14 Light source device 16 Processor device 18 Monitor 20 Input device 21 Insertion unit 22 Operation unit 22a Angle knob 22b Mode changeover switch 22c Zoom operation unit 23 Curved part 24 Tip part 24a, 24c Lighting Optical system 24b Imaging optical system 34 Blue laser light source 36 Blue purple laser light source 40 Light source control unit 41 Light guide 42a to 42d LED
43 Optical path coupling part 44 Fluorescent material 45 Illumination lens 46 Imaging lens 47 Zooming lens 48 Imaging sensor 50 CDS / AGC circuit 51 Gamma converter 52 A / D converter 54 Receiver 56 DSP
58 Noise removal unit 60 Image processing switching unit 62 Normal light image processing unit 64 Special light image processing unit 66 Image display signal generation unit 68 Image acquisition unit 70 Color enhancement unit 72 Structure enhancement unit 74 Image acquisition unit 77 Abnormal area enhancement unit 78 Structure Emphasis unit 80 Color space conversion unit 82 Color difference vector acquisition unit 84 Color difference expansion unit 86 Color space inverse conversion unit 88 Storage unit

Claims (15)

An image acquisition unit that acquires an endoscopic image composed of image signals including a narrow band image signal, and an image acquisition unit.
In a three-dimensional color space including brightness, a color difference vector which is a three-dimensional vector in the three-dimensional color space from the reference color of the endoscope image to the color of each attention pixel in the endoscope image is obtained. The color difference vector acquisition unit to be acquired and
In the color space, the color difference expansion vector of the vector amount obtained by adding the emphasis amount to the vector amount of the color difference vector in the same direction as the color difference vector of the attention pixel is added to the reference color. A color difference expansion unit that performs color difference expansion processing to convert to color,
It is provided with an output unit that outputs an endoscopic image in which the color difference between a normal portion and an abnormal portion is expanded by applying the color difference expansion process to all the pixels of the endoscope image.
The reference color is an image processing apparatus calculated from pixels of the endoscope image other than pixels having at least one color component exceeding a second threshold value among the color components of each pixel of the endoscope image. .. The image processing apparatus according to claim 1, wherein the color component exceeding the second threshold value is a color component deviating from the mucous membrane color. In a three-dimensional color space that includes the coordinate axes of the luminance components
The endoscopic image is further provided with an enhancement processing unit that performs structural enhancement processing such as spatial frequency enhancement.
The output unit outputs an endoscopic image obtained by applying the color difference expansion process and the structure enhancement process to the endoscopic image.
The color difference vector acquisition unit uses a reference color having a higher luminance value as the representative luminance value becomes higher, corresponding to a representative luminance value representing the endoscopic image among a plurality of predetermined reference colors. The image processing apparatus according to claim 1 or 2, wherein the color difference vector is acquired by using a reference color whose luminance value is lower as the representative luminance value becomes lower. The image processing apparatus according to any one of claims 1 to 3, wherein the reference color represents a color representing a normal portion. The image processing apparatus according to claim 4, wherein the color difference expanding unit expands the color difference between the reference color and a color appearing in the abnormal portion. The image processing apparatus according to any one of claims 1 to 5, wherein the color difference expanding portion increases the amount of color enhancement appearing in the abnormal portion to be larger than the amount of enhancement of colors other than the abnormal portion. When the vector amount of the color difference vector is smaller than the first threshold value, the color difference expansion unit increases the emphasis amount as the vector amount of the color difference vector increases, and the vector amount of the color difference vector is larger than the first threshold value. The image processing apparatus according to any one of claims 1 to 6, wherein when the amount is large, the amount of emphasis is reduced as the amount of vector increases. The image processing apparatus according to claim 7, wherein the first threshold value is a distance between a color representing the abnormal portion and the reference color. The image processing apparatus according to any one of claims 1 to 8, wherein the representative brightness value of the endoscope image is an average value of the brightness values of all the pixels of the endoscope image. The image according to any one of claims 1 to 9, wherein the color space is any one of Ycc color space, Lab color space, Luv color space, HSB color space, HSV color space, and HSV color space. Processing equipment. The image processing apparatus according to any one of claims 1 to 10, wherein the narrow-band image signal is obtained by imaging a sample illuminated with narrow-band light that absorbs more light into blood than other bands. The narrow band image signal is a blue narrow band image signal obtained by imaging a sample illuminated with blue narrow band light that absorbs more light into blood than other bands in the blue band, or blood in the green band. The image processing apparatus according to claim 11, which is a green narrow band image signal obtained by imaging a sample illuminated with green narrow band light that absorbs more light than other bands. The image processing apparatus according to any one of claims 1 to 12, wherein the endoscopic image is an image obtained by photographing the inner wall of the stomach, the normal portion is a normal mucous membrane, and the abnormal portion is an abnormal mucous membrane. It is an operation method of an image processing device including an image acquisition unit, a color difference vector acquisition unit, a color difference expansion unit, and an output unit.
An image acquisition step in which the image acquisition unit acquires an endoscopic image generated by an image signal including a narrow band image signal, and
In the three-dimensional color space including the brightness, the color difference vector acquisition unit is three-dimensional in the three-dimensional color space from the reference color of the endoscope image to the color of each attention pixel in the endoscope image. The color difference vector acquisition step to acquire the color difference vector, which is the vector of
In the color space, the color difference expansion unit obtains a color difference expansion vector having a vector amount in which the color of each attention pixel is in the same direction as the color difference vector of the attention pixel and the emphasis amount is added to the vector amount of the color difference vector. A color difference expansion step of performing a color difference expansion process for converting to a color added to the reference color, and
The output unit includes an output step of outputting an endoscope image in which the color difference between a normal portion and an abnormal portion is expanded by applying the color difference expansion process to all pixels of the endoscope image.
The reference color is an image processing apparatus calculated from pixels of the endoscope image other than pixels having at least one color component exceeding a second threshold among the color components of each pixel of the endoscope image. How to operate. Computer,
An image acquisition unit that acquires an endoscopic image generated by an image signal including a narrow band image signal, and an image acquisition unit.
In a three-dimensional color space including brightness, a color difference vector which is a three-dimensional vector in the three-dimensional color space from the reference color of the endoscope image to the color of each attention pixel in the endoscope image is obtained. The color difference vector acquisition unit to be acquired and
In the color space, the color difference expansion vector of the vector amount obtained by adding the emphasis amount to the vector amount of the color difference vector in the same direction as the color difference vector of the attention pixel is added to the reference color. A color difference expansion unit that performs color difference expansion processing to convert to color,
An image processing program for functioning as an output unit that outputs an endoscopic image in which the color difference between a normal portion and an abnormal portion is expanded by applying the color difference expansion processing to all pixels of the endoscope image.
The reference color is an image processing program calculated from pixels of the endoscope image other than pixels having at least one color component exceeding a second threshold value among the color components of each pixel of the endoscope image. ..

Images