JP5173130B2 - Electronic endoscope device - Google Patents

Description

  The present invention relates to an image processing function of an electronic endoscope apparatus, and more particularly to a function of adjusting a color.

  As one of the image processing functions of an endoscope apparatus, a function of enhancing a hue so that a difference between a normal site and a lesion site stands out is known. For example, Patent Document 1 discloses an apparatus that includes an operation switch for each observation region and each pigment application type, and recognizes the pressed switch to perform an optimum hue adjustment.

This apparatus sets a reference color phase by an emphasis color setting circuit, sets an optimum value of a hue width to be enlarged by an emphasis amount setting circuit, for example, ± 10 degrees, and sets the color set by the emphasis color setting circuit. A process of enlarging the surrounding hue width by the width set by the enhancement amount setting circuit using the phase as a reference is performed.
Japanese Patent Laid-Open No. 1-113018

  However, the method of expanding the hue by, for example, increasing the hue around the reference phase by ± 10 degrees only extends the hue around the reference phase linearly and cannot increase the hue nonlinearly. In order to improve the diagnosis, it is effective to make the color difference stand out, but there is also a need to eliminate the color difference unnecessary for the diagnosis.

  Accordingly, an object of the present invention is to provide an electronic endoscope image apparatus capable of performing not only linear hue expansion but also finer hue adjustment.

  The present invention provides an electronic endoscope apparatus having a function of adjusting the hue of an image acquired by an electronic endoscope, and having the following configuration.

  In this electronic endoscope apparatus, a value representing a color that can be acquired by the electronic endoscope by photographing an observation target and a hue in a color range necessary for diagnosis are wider than a hue in a color range unnecessary for diagnosis. A plurality of replacement tables are provided for each inspection purpose, in which associations with defined color values are stored. The replacement table is stored in a memory or the like of the electronic endoscope apparatus.

  “The hue of the color range necessary for diagnosis is wider than the hue of the color range unnecessary for diagnosis” means that the hue of the color range necessary for diagnosis is expanded and the hue of the color range unnecessary for diagnosis does not change. Or the hue of the color range necessary for diagnosis does not change and the hue of the color range unnecessary for diagnosis is reduced, or the hue of the color range necessary for diagnosis is expanded and the hue of the color range unnecessary for diagnosis Is reduced.

  “Necessary for diagnosis” means that a doctor as a user may make some diagnosis based on the fact that a color difference is observed in the color range. “Unnecessary for diagnosis” means that even if there is a color difference in the color range, the diagnosis result is not affected.

  In addition, the electronic endoscope apparatus uses the correspondence stored in the replacement table for one examination purpose designated by the input instruction data, and configures each image constituting the image acquired by the electronic endoscope. A purpose-specific image generation means for generating an image for the selected inspection purpose by individually replacing the color value of the pixel is provided.

  Since the electronic endoscope apparatus of the present invention adjusts the hue of an image by replacing the color value of each pixel constituting the image using a replacement table, the replacement table should be designed for the purpose of inspection. Therefore, it is possible to obtain a highly diagnostic image with finely adjusted hue.

  Hereinafter, as one embodiment of the present invention, an electronic endoscope system used for a digestive organ examination will be exemplified.

1. System Configuration FIG. 1 shows a schematic configuration of an electronic endoscope system. As shown in the figure, the electronic endoscope system 1 includes an electronic endoscope 2 (hereinafter referred to as a scope 2), a processing device 3 (hereinafter referred to as a processor 3) for processing an image acquired by the electronic endoscope, A light source device, a monitor, a printer, and the like (not shown) are provided. The electronic endoscope system 1 can use a plurality of types of scopes according to the purpose of the endoscopy, but the scope 2 in the figure shows a configuration common to these scopes.

  The scope 2 includes a CCD (Charge Coupled Device) 21, a signal processing circuit 22 that processes a signal acquired by the CCD 21, a microcomputer 23 that performs various control processes, and a connector unit (not shown) connected to the processor 3.

  The CCD 21 is attached to the tip of the scope 2 together with the objective lens, acquires reflected light from the observation target, and converts it into an electrical signal. In this embodiment, the imaging resolution of the CCD is about 5 μm.

  The signal processing circuit 22 performs signal processing such as correlated double sampling, automatic gain control, and A / D conversion on the output signal of the CCD 21. The microcomputer 23 controls the operation of the signal processing circuit and data transmission to the processor 3.

  Here, the acquisition method of the color information of the scope 2 will be described. In general, a CCD includes a color filter, and acquires information about the color of the observation target by acquiring reflected light from the observation target through the color filter. For this reason, the quality and color representation of the color information of the acquired image differ depending on the arrangement of color filters provided in the CCD or the type of color filter.

  For example, regarding the arrangement of the color filter, a frame sequential CCD that alternately acquires information of each color while rotating the rotary filter, and a simultaneous CCD that collectively acquires information of each color via the mosaic filter It has been known. Some existing scopes employ a field sequential CCD, while others employ a simultaneous CCD. However, in the case of the frame sequential type CCD, when the movement of the part to be imaged is faster than the rotation of the filter, the images of the respective colors may shift. For this reason, the electronic endoscope system 1 employs a scope having a simultaneous CCD as the scope 2.

  As the types of color filters, primary color filters for separating three colors of red (R), green (G), and blue (B), cyan (C), magenta (M), yellow (Y), green ( A complementary color filter that separates the four colors G) is known. An image acquired by a CCD provided with a primary color filter (hereinafter referred to as primary color CCD) is expressed in RGB, and an image acquired by a CCD provided with a complementary color filter (hereinafter referred to as complementary color CCD) is expressed in CMYG. For this reason, the existing electronic endoscope system can use only one of a scope having a primary color CCD and a scope having a complementary color CCD. On the other hand, in the electronic endoscope system 1, a plurality of types of scopes having different types of CCD color filters can be used.

Table 1 illustrates scopes that can be used in the electronic endoscope system 1.

  In Table 1, the CCDs of scopes A, B, and C are all CCDs having primary color filters and have the same resolution. However, scopes B and C have a specific color light separately from the CCD filters. It has a cut filter that prevents it from passing through. As described above, the method for acquiring the color information of the observation target is not only the field sequential type or the simultaneous type, the primary color CCD or the complementary color CCD, but also the type of cut filter combined with the CCD.

  Next, the configuration of the processor 3 will be described. The processor 3 includes a connector unit (not shown). The connector part of the processor 3 has a structure in which the connector part of each scope can be easily connected or removed.

  The processor 3 includes a signal processing circuit 31 that performs gamma correction on a signal input from the signal processing circuit 22 of the scope 2 via the connector unit to generate a video signal. When the output signal of the scope signal processing circuit 22 is a CMYG signal, the signal processing circuit 31 also performs a conversion process from the CMYG signal to the RGB signal. The processor 3 includes a microcomputer 32 that controls the operation of the signal processing circuit 31 and the communication with the scope 2. A signal processing circuit 35 that generates a monitor output signal by performing pixel number conversion and D / A conversion is disposed at the subsequent stage of the signal processing circuit 31.

  Furthermore, the processor 3 includes an image processing dedicated substrate 4 separately from the main substrate on which the signal processing circuit 31, the microcomputer 32, and the signal processing circuit 35 are mounted. On the image processing dedicated board 4, an image processing circuit 41 that performs various image processing on the image signal output from the signal processing circuit 31 and a microcomputer 42 that controls the image processing circuit 41 are mounted. The image processing circuit 41 is connected to the signal processing circuits 31 and 35 via the selectors 33 and 34. The selectors 33 and 34 are switched based on a control signal from the microcomputer 32.

  FIG. 2 shows a detailed configuration of the image processing dedicated substrate 4. As shown in the figure, the image processing circuit 41 is classified into six processing units: a standard image generation unit 411, a color replacement unit 412, a lightness / saturation adjustment unit 413, a sharpness adjustment unit 414, a hue adjustment unit 415, and a post-processing unit 416. Is done. The color replacement unit 412, the brightness / saturation adjustment unit 413, the sharpness adjustment unit 414, and the hue adjustment unit 415 can be selectively operated by switching the selectors 417a to 417e. The selectors 417a to 417e are switched based on a control signal supplied from the microcomputer 42.

  Further, the memory 43 is mounted on the image processing dedicated substrate 4. The memory 43 stores various look-up tables used in the processing of the color replacement unit 412, the brightness / saturation adjustment unit 413, the sharpness adjustment unit 414, and the hue adjustment unit 415. Among these, the lookup table used in the color replacement unit 412 is stored for each scope model. In addition to this, the memory 43 stores various processing parameters. Note that the look-up table and processing parameters stored in the memory 43 are not only stored when the electronic endoscope system 1 is shipped from the manufacturer, but also stored by the user later. include. These lookup tables and parameters are referred to by the microcomputer 42 as described later.

  Here, the switching between the selectors 33 and 34 and the selectors 417a to 417e will be further described. Whether or not to use the function of the image processing board 4 is set on the operation panel. Whether to use the functions of the color replacement unit 412, the brightness / saturation adjustment unit 413, the sharpness adjustment unit 414, and the hue adjustment unit 415 when using the function of the image processing dedicated substrate 4 is also set on the operation panel. The Information on settings made by the user on the operation panel (hereinafter, setting information) is stored in the memory. This memory is preferably an independent memory for storing setting information. However, a partial area of the memory 43 or a memory provided for other purposes may be used.

  This memory can store a plurality of sets of setting information related to function selection. Therefore, each time the user uses the electronic endoscope system, the user can select a function on the operation panel, or the setting information is stored in advance for each inspection purpose. Information can also be recalled and used.

  For example, setting information suitable for observation of the stomach, setting suitable for observation of the large intestine may be stored for each observation site, setting suitable for examination using a dark blue test agent, reddish brown You may memorize | store setting information for every test agent to be used like the setting suitable for the test | inspection using a test agent. Furthermore, it may be stored for each object to be discovered by examination such as a setting suitable for finding an aneurysm and a setting suitable for finding a tumor.

  FIG. 3 is a flowchart showing an overview of the initialization process executed by the microcomputer 32 when the power is turned on. When the scope 2 is connected to the processor 3 and the power of the electronic endoscope system 1 is turned on, the microcomputer 32 communicates with the microcomputer 23 of the scope 2 to determine the data for determining the model of the scope 2 (hereinafter referred to as “the scope 2”). (Referred to as scope data) is acquired (S101). The scope data includes information illustrated in Table 1, that is, information indicating the resolution of the CCD and the type of filter. The information indicating the type of filter includes information such as primary / complementary color distinction, presence / absence of a cut filter, and a frequency (color) cut by the cut filter.

  Subsequently, the microcomputer 32 refers to the setting information stored in the memory (S102). When a plurality of setting information is stored, the setting information selected by the user by operating the operation panel is referred to. When the user performs an operation of selecting setting information on the operation panel, a selection signal representing the selection content is input to the microcomputer 32. The microcomputer 32 determines the setting information selected based on the input selection signal. Then, it is determined whether connection with the image processing board 4 is necessary or not based on the reference setting information (S103), and the selectors 33 and 34 are controlled based on the determination result (S104 and S109). When the connection is unnecessary and the image processing board 4 is disconnected, the initialization process is terminated at that time.

  When the microcomputer 32 controls the selector to connect the image processing dedicated substrate 4 in step S104, the microcomputer 32 determines whether or not the combination of functions selected by the user is appropriate based on the setting information referenced in step S102. Determination is made (S105).

  If the microcomputer 32 determines that the combination of the selection functions is appropriate, the microcomputer 32 sends instruction data representing the selected contents to the microcomputer 42 of the image processing board 4 together with the scope data acquired in step S101 (S108). .

  On the other hand, if it is determined that the combination of the selection functions is not appropriate, a warning message is output to a monitor device (not shown) connected to the processor 3 (S106). In this case, the microcomputer 32 automatically corrects the combination of functions to an appropriate combination (S107), and sends instruction data representing the content to the microcomputer 42 of the image processing board 4 together with the scope data acquired in step S101. (S108).

  Note that after outputting the warning message, the scope data and the instruction data may be transmitted after waiting for the user to perform a confirmation operation.

  The determination as to whether the function selection is appropriate is performed based on a determination rule stored in advance. In a memory provided in the processor 3, for example, data defining a determination rule that the color replacement unit must be operated whenever the lightness / saturation adjustment unit is operated is stored. Then, by checking the setting information with the data, it is determined whether the function selection performed by the user is appropriate.

  Next, the operation of the microcomputer 42 will be described with reference to FIG. When receiving the scope data and the instruction data sent from the microcomputer 32 (S201), the microcomputer 42 determines the scope model based on the scope data, and further determines the function selected based on the instruction data (S202). Next, the microcomputer 42 controls the selectors 417a to 417e so that the signal is input to the processing unit that provides the selected function (S203). Subsequently, the microcomputer 42 reads a lookup table and parameters required by each processing unit from the memory 43 (S204).

  At this time, as described above, the lookup table includes a table stored for each scope model. The scope model discrimination result is used when selecting a lookup table to be read from the memory. The read lookup table and parameters are sent to a processing unit that requires them (S205).

  As described above, the processor 3 of the electronic endoscope system 1 performs the image processing after determining the type of the scope 2, in other words, the color information acquisition method of the scope, and therefore adopts any type of scope as the scope 2. can do. That is, the present invention is not limited to a type having a complementary color CCD or a type having a primary color CCD.

  The user can freely select an image processing function to be used by performing a predetermined setting operation on the operation panel or by calling a setting stored in advance. At this time, even if an inappropriate function is selected due to lack of knowledge or the like, the appropriate selection is automatically performed together with the output of the warning message, so that the image processing function can always be used effectively.

  Furthermore, a user who does not use the function of the image processing dedicated substrate 4 can stop signal input to the image processing dedicated substrate 4 by performing a predetermined setting operation on the operation panel. As a result, power is not consumed for functions that are not used.

  The image processing circuit 41 includes a plurality of functions that individually provide the functions of the standard image generation unit 411, the color replacement unit 412, the lightness / saturation adjustment unit 413, the sharpness adjustment unit 414, the hue adjustment unit 415, and the post-processing unit 416. A circuit in which a semiconductor device is arranged may be used, or a circuit in which a CPU dedicated to image processing and a memory in which six types of image processing programs are stored are arranged, and whether or not each process is executed in the program. It may be one that switches between them.

2. Image Processing Function of System Hereinafter, image processing executed by the image processing circuit 41 will be described in detail. The image processing functions provided by the image processing circuit 41 are roughly classified into two types.

  The first function is a function that absorbs the difference in the scope color information acquisition method. Specifically, the difference in the color information acquisition method is absorbed by generating a standard image that does not depend on the scope model, specifically, the type of CCD color filter. The standard image is an image in which the color of the imaged part is faithfully reproduced. This standard image is generated by the standard image generation unit 411.

  The second function is a function of processing the standard image generated by the first function so as to become an image suitable for diagnosis. This function is provided by the color replacement unit 412, the brightness / saturation adjustment unit 413, the sharpness adjustment unit 414, and the hue adjustment unit 415.

  The color replacement unit 412 performs image processing for the purpose of matching the color of the image to the user's preference. The lightness / saturation adjusting unit 413 performs image processing for the purpose of making it easy to see a dark and difficult-to-see part. The sharpness adjustment unit 414 performs image processing for the purpose of enhancing structures (for example, unevenness and blood vessels) necessary for diagnosis. In addition, the hue adjustment unit 415 performs image processing for emphasizing the color difference necessary for diagnosis and reducing the color difference that is unnecessary for diagnosis or interferes with diagnosis. The post-processing unit 416 performs noise removal and monitor output signal generation processing.

  Hereinafter, specific examples of processing are shown for each unit.

2.1 Standard Image Generation Unit The standard image generation unit 411 is a processing unit that generates a standard image, and performs predetermined preprocessing before generating a standard image. First, the preprocessing will be described.

  As described above with reference to the signal processing circuit 31, in the electronic endoscope system 1, when the scope 2 is a scope having a complementary color CCD, the signal processing circuit 31 converts the CMYG signal into the RGB signal. . More specifically, the signal processing circuit 31 converts a signal obtained from the complementary color CCD into a luminance signal Y and color difference signals Cr and Cb, and further uses the luminance signal Y and the color difference signals Cr and Cb (hereinafter referred to as YCC signals) as primary colors. Convert to signals R, G, B.

  For this reason, each value of the RGB signal input to the image processing circuit 41 may be a value including a value after the decimal point. If such a value is used for a complicated calculation, a rounding error may occur in the process of repeating the calculation for image processing, and there is a possibility that the color space will be distorted.

  In view of this, the standard image generation unit 411 first executes a process of replacing 8-bit data with 10-bit data. Specifically, for each of the R value, G value, and B value of all the pixels constituting the image, a 10-bit area is secured, and a value obtained by quadrupling 8-bit data is stored in the area. As shown in FIG. 5, this is nothing but storing a value obtained by shifting an 8-bit value by 2 bits in the upper direction in a 10-bit area and storing a 0 value in the lower 2 bits of the area. . As a result, the number of significant digits in the calculation increases, so that the influence of errors caused by rounding can be reduced. Note that the data converted into 10 bits and processed by the standard image generation unit 411 is converted again into 8-bit data by the post-processing unit 416 and then output to the monitor.

  Next, a process for generating a standard image using the 10-bit data will be described.

  The standard image is generated by converting (replacing) the 10-bit RGB data using a previously created three-dimensional lookup table 5 (hereinafter, three-dimensional LUT 5).

  As shown in FIG. 6, the three-dimensional LUT 5 is a table that replaces color values (R, G, B) acquired by the scope with real color values (R ′, B ′, G ′). The color value that may be acquired by the scope can be acquired by actually observing each part using the scope. In addition, the actual color value can be obtained by measuring the color of each part using a colorimetric system using the opportunity of surgery. Therefore, if data acquired from the same part are associated with each other, the three-dimensional LUT 5 for the scope can be created.

  If data is collected by observing each scope model, a three-dimensional LUT 5 for each model can be created. According to the LUT 5 created by the above procedure, not only the difference in the color information acquisition method of the scope but also the difference related to other than the color such as the difference in the aberration of the objective lens can be absorbed.

  The three-dimensional LUT 5 for each model created in this way is stored in the memory 43 of FIG. 2, and is supplied from the microcomputer 42 to the standard image generation unit 411 according to the procedure described with reference to FIG. The standard image generation unit 411 generates a standard image that faithfully reproduces the true color of the observation target by replacing the R, G, and B values of each pixel constituting the image using the three-dimensional LUT 5.

The three-dimensional LUT 5 that replaces data in which each R, G, B value is represented by 10 bits with data in which each R, G, B value is represented by 10 bits similarly stores 1,024 ³ associations. Will do. Further, since it is necessary to store a plurality of three-dimensional LUTs 5 for each scope model, the memory 43 needs to have a capacity sufficient to store them.

  As a countermeasure when the memory capacity is limited, a method of storing only a part of associations in the LUT and obtaining other associations by interpolation calculation is conceivable. For example, as illustrated in FIG. 7, only the association when R, G, and B are 0, 32, 63, 96,..., 255 (32 values) is stored in the three-dimensional LUT 6. deep.

  When an RGB value that is not stored in the three-dimensional LUT 6 is input, as shown in FIG. 8, the three-dimensional LUT 6 starts from the periphery of the coordinate point P (r, g, b) in the RGB space of the input value. The coordinate points A to H for which the correspondence is stored are extracted. Then, the RGB values of the coordinate points A to H are replaced with the R′G′B ′ values of the coordinate points A ′ to H ′ using the three-dimensional LUT 6.

  After that, if the R′G′B ′ values of the coordinate points A ′ to H ′ are weighted according to the relationship between the coordinate points A to H and the coordinate point P, the association is stored in the three-dimensional LUT 6. P ′ (r ′, g ′, b ′) corresponding to the coordinate point P (r, g, b) that is not present can be obtained.

  Through the processing described above, it is possible to obtain a standard image in which a color faithful to the actual color is reproduced without depending on the scope model. Since the subsequent image processing is performed on this standard image, there is no difference in the finish of the image output by the image processing circuit 41 depending on the scope model.

  In the present embodiment, the standard image generated by the standard image generation unit 411 can be output to the monitor without switching the selectors 417a to 417e so that the processing unit subsequent to the color replacement unit is not executed.

2.2 Color Replacement Unit The color replacement unit 412 performs color replacement processing for the purpose of matching the color of an image to the user's preference.

  For example, a xenon lamp is often used as a light source of an electronic endoscope system, but since an early endoscope used a halogen lamp as a light source, an acquired image was yellowish. For this reason, there are many users who think that an image having a familiar color can be diagnosed more accurately than an image in which the actual color is faithfully reproduced. In order to meet such needs, the color replacement unit 412 replaces the color of the standard image with a user's favorite color.

  The color replacement is performed using a three-dimensional LUT, as in the standard image generation process. The three-dimensional LUT used here is a table that replaces the RGB values of the standard image with R′G′B ′ values that the user likes. This table may be provided by the manufacturer that provides the electronic endoscope system 1 or may be created by the user himself / herself.

  The three-dimensional LUT used by the color replacement unit 412 is also stored in the memory 43 of FIG. 2, and is supplied from the microcomputer 42 to the color replacement unit 412 in the procedure described with reference to FIG. The setting information includes information necessary for table selection. When a plurality of three-dimensional LUTs are stored, the table selected based on the setting information is supplied to the color replacement unit 412. The color replacement unit 412 converts the RGB value of each pixel constituting the standard image using the supplied three-dimensional LUT.

  Similar to the standard image generation processing, only a part of the association may be stored in the three-dimensional LUT, and the conversion of the RGB values for which the association is not stored in the three-dimensional LUT may be performed by interpolation calculation.

2.3 Brightness / Saturation Adjustment Unit The brightness / saturation adjustment unit 413 performs processing for adjusting the brightness of an image. In particular, image processing is performed for the purpose of solving a problem caused by a large difference in brightness in an image.

  For example, in endoscopic imaging of the entrance portion of the duodenum, an image with a large contrast between the front (stomach side) and the back (in the duodenum) is obtained. In particular, when photographing is performed by spraying a dark blue test drug such as dark blue, the inside of the duodenum is too dark to be observed. For diagnosis, it is necessary to have an image that can be seen from the front to the back even when sprayed with test chemicals. Unlike conventional photography, the position and orientation of the illumination are adjusted when photographing with an endoscope. I can't do it. Also, there is a risk of burns to the human body, so the lighting cannot be increased. For this reason, it is necessary to adjust by image processing so that the brightness of the image becomes an appropriate brightness.

  At this time, since the CCD is an element that controls the output voltage in accordance with the amount of light, the output voltage becomes small when photographing in a dark region, and color information may not be obtained accurately. In this case, if the value of the luminance component of the image is simply increased by image processing, there may occur a problem that dark blue does not become light blue but becomes green. For this reason, in the process of the lightness / saturation adjustment unit 413, the brightness is adjusted while adjusting the color rather than simply correcting the lightness of the image.

  FIG. 9 is a flowchart showing an outline of processing of the lightness / saturation adjusting unit 413. The lightness / saturation adjustment unit 413 first converts the input image into a YCC signal, and generates a luminance image Y consisting only of luminance components (S301). In the following description, the position of each pixel constituting the image is represented by (x, y), and the value of the pixel at the position (x, y) of the image I is represented by I (x, y). .

  Subsequently, the brightness image YY is generated by performing a blurring process using an image processing filter on the brightness image Y. The blurring process is performed in two stages. In the first stage blurring process, a moving average filter is used as an image processing filter (S302). In the second stage blurring process, a Gaussian filter is used as an image processing filter. By this two-stage blurring process, a luminance blur image UY representing the luminance distribution of the image is generated (S303).

  Next, the lightness / saturation adjustment unit 413 determines a brightness adjustment amount indicating the degree of brightness adjustment (S304). The brightness adjustment is determined for each pixel by converting each pixel value of the luminance blurred image UY using the lookup table LUTy.

  The LUTy is a one-dimensional lookup table that stores associations between input values in the range of 0 to 1023 and output values in the range of 0 to 1023. As the LUTy, for example, as shown in FIG. 10, a table is used in which the output value becomes 0 when the input value exceeds a predetermined value. FIG. 10 represents the association between the input value and the output value with the input value as the horizontal axis and the output value as the vertical axis.

  The LUTy may be appropriately designed based on the brightness adjustment policy. For example, the table illustrated in FIG. 10 is a table designed based on an adjustment policy of brightening a dark part and leaving a bright part as it is. However, the dark part is significantly brightened and the bright part is slightly brightened. Alternatively, various adjustment policies are conceivable, such as slightly reducing the brightness of bright areas. Note that the LUTy is stored in the memory 43 of FIG. 2, and is supplied from the microcomputer 42 to the lightness / saturation adjustment unit 413 in the procedure described with reference to FIG.

Next, distribution of brightness contribution and saturation contribution in the brightness adjustment is determined (S305). This distribution is determined for each pixel constituting the image. Specifically, the lightness contribution rate 1 and the saturation contribution rate 2 are determined by the following equations.
Yp (x, y) = Y (x, y) + LUTy (UY (x, y))
rate1 (x, y) = LUTy (UY (x, y)) × LUTr (Yp)
rate2 (x, y) = LUTy (UY (x, y)) × (1-LUTr (Yp))
However, Yp is a value obtained by estimating the luminance value of each pixel when adjustment is performed based only on luminance information obtained from the luminance blurred image UY.

  For example, as shown in FIG. 11, the output value of the LUTr is almost constant when the input value is equal to or lower than the predetermined value and when the input value is equal to or higher than the predetermined value. Increase the table. This table is a table designed based on a policy that the saturation adjustment function is particularly effective in a dark region. The LUTr may be appropriately designed based on the adjustment policy, similar to the LUTy. The LUTr is also stored in the memory 43 of FIG. 2, and is supplied from the microcomputer 42 to the lightness / saturation adjustment unit 413 in the procedure described with reference to FIG.

  As is apparent from the above equation, the sum of rate1 (x, y) and rate2 (x, y) is the brightness adjustment amount LUTy (UY (x, y)) determined in step S304. It is determined. The distribution of rate1 (x, y) and rate2 (x, y) depends on the estimated luminance value. This is to determine how much the brightness is adjusted by conversion by LUTy, and then to determine how much the saturation adjustment is to be applied within the determined amount. Furthermore, when deciding how much saturation adjustment should be applied, the estimated value of what happens if only the brightness is adjusted without adjusting the saturation is referred to.

Next, the lightness / saturation adjustment unit 413 performs saturation on the RGB values of each pixel (x, y) constituting the input signal according to the following formula, by weighting the RGB values according to the saturation contribution. The adjustment value and the brightness adjustment value weighted according to the brightness contribution to the pixel value of the luminance image are added (S306).
R ′ (x, y) = R (x, y) + Y (x, y) × rate1 + R (x, y) × rate2
G ′ (x, y) = G (x, y) + Y (x, y) × rate1 + G (x, y) × rate2
B ′ (x, y) = B (x, y) + Y (x, y) × rate1 + B (x, y) × rate2

Alternatively, the lightness adjustment value may be a value obtained by weighting the pixel value of the luminance blurred image according to the lightness contribution. That is, in step S306, processing according to the following equation may be executed.
R ′ (x, y) = R (x, y) + UY (x, y) × rate1 + R (x, y) × rate2
G ′ (x, y) = G (x, y) + UY (x, y) × rate1 + G (x, y) × rate2
B ′ (x, y) = B (x, y) + UY (x, y) × rate1 + B (x, y) × rate2

  In the brightness adjustment process using the above processing formula, not only the brightness is adjusted, but also the saturation is adjusted, so that an image with a natural hue and a little uncomfortable feeling can be obtained. For example, in the process using the look-up table illustrated in FIGS. 10 and 11, particularly careful saturation adjustment is performed in a dark part, so that even a fine structure in a dark part can be reproduced easily. In addition, for example, a dark part is unnecessary for diagnosis and can be left as it is, and a look-up table can be designed so that a fine structure can be easily reproduced in an area necessary for diagnosis. That is, the brightness can be adjusted by an amount suitable for the purpose by a method suitable for the purpose.

  Here, the blurring process in steps S302 and S303 will be further described.

  A method for generating a blurred image of an image by replacing the value of each pixel constituting the image with a value obtained by calculation using an image processing filter has been conventionally known. In such a blurred image generation process, an image processing filter having a size of about 3 × 3 pixels to 15 × 15 pixels is usually used for calculation. On the other hand, the lightness / saturation adjusting unit 413 performs an operation using an image processing filter having a width equal to or larger than 1/4 of the width of the image. For example, when the width of the image is 1024 pixels, an image processing filter having a size of 255 × 255 pixels or larger is used.

  FIG. 12 is a diagram for explaining the relationship between the size of the image processing filter and the processing result. FIG. 12A shows an example of normal image processing, and FIG. 12B shows an example of image processing of the lightness / saturation adjustment unit 413.

When a 3 × 3 pixel image processing filter is used as illustrated in FIG. 12A, the value of each pixel depends on the value of nine pixels. On the other hand, in the case of using the image processing filter 255 × 255 pixels as shown in FIG. 12 (b), the value of each pixel, will depend on the value of the 255 ^two pixels.

  When adjusting the brightness of an image with a large difference in brightness, it is not just that the dark part needs to be brightened, but the difference in brightness must be left to the extent that it does not become unnatural. That is, it is necessary to brighten the dark part without breaking the balance of the brightness of the entire image. For this purpose, it is necessary to determine the brightness based on the values of distant pixels rather than determining the brightness only by looking at the values of adjacent pixels.

  For example, each pixel value of a luminance blurred image generated using a small image processing filter as illustrated in FIG. 12A is determined regardless of the values of pixels that are separated by 100 pixels or more. For this reason, it cannot be said that there is no possibility that the lightness and darkness is reversed between the pixels separated by 100 pixels or more.

  On the other hand, each pixel value of a luminance blurred image generated using a large image processing filter as illustrated in FIG. 12A is determined with reference to the values of pixels separated by 100 pixels or more. Therefore, it is impossible for light and dark to be reversed between pixels that are separated by 100 pixels or more, and the balance of light and dark of the entire image will not be lost.

  In the process of the lightness / saturation adjustment unit 413, the brightness blur image UY that faithfully reflects the balance of lightness and darkness of the entire image is used for calculation of the estimated brightness and determination of the lightness contribution and saturation contribution. Therefore, the image after adjustment is a natural image in which the balance between light and dark is maintained.

  Here, if the size of the image processing filter is large, as a matter of course, the amount of calculation increases. However, since the image processing of the electronic endoscope requires real-time characteristics unlike the image processing of the X-ray apparatus, it is not preferable that the processing time becomes long.

  Therefore, the lightness / saturation adjusting unit 413 reduces the amount of calculation by performing the calculation by arranging filters every several pixels in the vertical and horizontal directions when performing the first blurring process in step S302. For example, if the calculation is executed every three pixels, the calculation amount can be reduced to 1/9 that of the calculation for all the pixels, and the processing time can be shortened.

  In the second stage blurring process in step S303, the calculation may be performed every several pixels in the same manner. However, in the calculation using the Gaussian filter, an artifact may occur in the calculation every several pixels. The first-stage blurring process is preferably performed for every several pixels, and the second-stage blurring process is performed for all pixels.

2.4 Sharpness Adjustment Unit The sharpness adjustment unit 414 performs image processing mainly for the purpose of enhancing the sharpness of an image. However, since the endoscopic image does not include a sharp edge, processing mainly for enhancing the structure of the observation target such as unevenness and blood vessels is performed.

  For example, if any protuberance is formed under the stomach mucosa, the thin blood vessels are spiraled and the thick blood vessels are curved. For this reason, the observation of the change in the shape of the blood vessel is extremely effective in finding a hidden lesion. However, since the inside of the stomach has a similar color as a whole, when trying to distinguish blood vessels based on the frequency components of the image, it is not possible to distinguish between thick blood vessels and shadows of raised objects.

  For this reason, the sharpness adjustment unit 414 executes a sharpness adjustment process in consideration of the color of the image.

  FIG. 13 is a flowchart showing an outline of processing of the sharpness adjustment unit 414.

  The sharpness adjustment unit 414 first converts the input image into a YCC signal, and generates a luminance image Y (x, y) consisting only of luminance components (S401).

  Subsequently, the luminance image Y (x, y) is subjected to a blurring process using an image processing filter to generate a luminance blurred image U (Y (x, y)). The blurring process is performed in two stages. In the first stage blurring process, a moving average filter is used as an image processing filter (S402). If the filter is too small, information of a relatively large structure such as the unevenness of the stomach wall is not included, so the size is about 80 × 80 pixels. In step S402, similarly to step S302 of the lightness / saturation adjustment unit 413, calculation is executed every several pixels so that the calculation amount is reduced.

  In the second-stage blurring process, a Gaussian filter is used as an image processing filter (S403). The Gaussian filter uses three types of filters with different standard deviations.

  14A, 14B and 14C show three types of Gaussian filters. In the figure, the horizontal axis represents the position of each pixel with the position of the center pixel of the image processing filter being 0, and the vertical axis represents the value of each pixel constituting the filter. FIG. 14A shows a filter having the largest standard deviation among the three types having a size of 79 × 79 pixels. With this filter, information on relatively large structures such as irregularities on the inner wall of the organ and arteries is extracted. FIG. 14B is a filter having a size of about 15 × 15 pixels and a standard deviation smaller than that of the filter of FIG. This filter extracts medium-sized structure information such as a standard-sized blood vessel. FIG. 14C shows a filter having a size of about 3 × 3 pixels and a smaller standard deviation than the filter shown in FIG. This filter extracts information on fine structures such as capillaries.

  In step S403, if a Gaussian filter having a standard deviation set according to the size of the structure to be enhanced is used, a luminance blurred image UY in which image information sufficient to distinguish the structure to be enhanced is left is generated. Can do.

  In the second-stage blurring process, three types of computations may be performed using three types of Gaussian filters. However, in order to shorten the processing time, three types of Gaussian filters are superimposed (function It is preferable to perform the calculation using an image processing filter (also provided).

When the sharpness adjustment unit 414 generates the brightness blur image UY, the sharpness enhancement amount C is calculated by calculating the difference between the values of the brightness image Y and the brightness blur image UY for each pixel by the following equation. (X, y) is determined (step S404).
C (x, y) = | Y (x, y) −UY (x, y) | × LUTy (UY (x, y))
The lookup table LUTy may be the same table as the lookup table LUTy used by the lightness / saturation adjustment unit 413, or may be a table designed based on another adjustment policy.

  In general sharpness enhancement processing, sharpness is enhanced by simply adding the sharpness enhancement amount C (x, y) to each pixel value of an input image. On the other hand, in the processing of the sharpness adjustment unit 414, the color-dependent enhancement amounts Cr (x, y), Cg (x, y), and Cb (x, y) depending on the color are obtained (S405), and the color-specific enhancement is performed. Sharpness is enhanced by adding the amount to each pixel value of the input image.

The amount of enhancement for each color is calculated by the following procedure. First, based on the following equation, a lightness contribution rate 3 representing the contribution of lightness adjustment in sharpness enhancement and a saturation contribution rate 4 representing the contribution of saturation adjustment in sharpness enhancement are determined.
rate3 = LUTc (Y (x, y)) × C (x, y)
rate4 = {1-LUTc (Y (x, y))} × C (x, y)

  As is clear from the above equation, the lightness contribution and the saturation contribution are determined so that the sum thereof is the sharpness enhancement amount obtained in step S404. The LUTc is a table that determines the distribution of lightness contribution and saturation contribution, and can be appropriately defined according to the target image quality.

Subsequently, the emphasis amount for each color is determined by the following equation.
Cr (x, y) = Y (x, y) × rate3 + R (x, y) × rate4
Cg (x, y) = Y (x, y) × rate3 + G (x, y) × rate4
Cb (x, y) = Y (x, y) × rate3 + B (x, y) × rate4

Alternatively, the color enhancement amount may be determined using the following equation.
Cr (x, y) = C (x, y) × R (x, y) / Y (x, y)
Cg (x, y) = C (x, y) × G (x, y) / Y (x, y)
Cb (x, y) = C (x, y) × B (x, y) / Y (x, y)

Next, the sharpness is enhanced by adding the color-specific enhancement amount calculated using either of the above two formulas to the input signal as shown in the following formula (S406).
R ′ (x, y) = R (x, y) + Cr (x, y)
G ′ (x, y) = G (x, y) + Cg (x, y)
B ′ (x, y) = B (x, y) + Cb (x, y)

  If sharpness is enhanced by adding the amount of enhancement for each color, it is possible to realize sharpness enhancement according to the color. In this case, since the emphasized color component is different between the thick blood vessel and the shadow of the raised object, the shadow is not mistaken for the blood vessel.

  As is clear from the above description, the brightness / saturation adjustment unit 413 and the sharpness adjustment unit 414 both generate a luminance blurred image and use it for processing. Therefore, a processing unit that generates a luminance blur image is provided in the preceding stage of the selector 417a in FIG. 2, and the luminance blur image generated by the processing unit is input to the lightness / saturation adjustment unit 413 and the sharpness adjustment unit 414. It may be. In this case, the image processing filter used for generating the luminance blurred image is a filter having two types of functions.

  Note that the look-up table and the image processing filter used by the sharpness adjustment unit 414 are stored in the memory 43 of FIG. 2, and are supplied from the microcomputer 42 to the sharpness adjustment unit 414 in the procedure described with reference to FIG. .

2.5 Hue Adjustment Unit The hue adjustment unit 415 performs image processing for enlarging a hue necessary for diagnosis and reducing a hue unnecessary for diagnosis.

  In endoscopy, observation may be performed by spraying a test drug containing a pigment. For example, when a deep blue test drug called indigo carmine is sprayed on the inner wall of the stomach, the test drug accumulates in the valleys of the mucosal surface, so that the unevenness of the stomach wall can be observed as a contrast between red and blue colors. In addition, when spraying a deep blue test agent called methylene blue, the normal mucous membrane is stained blue, but the tumor is not stained. Therefore, the presence and location of the tumor can be confirmed.

  On the other hand, in such examinations, information that the skin tone is slightly different for each part is not necessary for diagnosis.

  FIG. 15 is a flowchart showing an outline of processing of the hue adjustment unit 415. The hue adjustment unit 415 first generates a hue image representing a hue distribution by converting the RGB signal into an HSV signal and extracting only the hue component (S501). Subsequently, the value H (x, y) of each pixel constituting the hue image is converted using the lookup table LUTh (S502).

  The LUTh is a table designed so that the hue is expanded in the color range necessary for diagnosis and the hue is narrowed in the color range unnecessary for diagnosis.

  For example, according to LUTh illustrated in FIG. 16, blue-violet is replaced with a color closer to blue. If the image is processed using this LUTh, the color contrast between the part where the test drug is attached and the part of the skin where the test drug is not attached will be increased when shooting is performed by spraying a blue-violet test drug. Can do. Further, according to the LUTh in FIG. 16, all the colors from red to pink are replaced with the same pink. For this reason, if an image is processed using this LUTh, a subtle difference in skin redness does not appear in the image.

  The LUTh illustrated in FIG. 17 is a table in which only the color of the hemoglobin pigment is replaced with green and other colors are not replaced. This table is designed for the purpose of generating an image suitable for blood flow observation by enlarging the difference between the color of blood and the color of skin.

  In the endoscopic image, the range of colors acquired by photographing is limited to a reddish color, and a blue or green target cannot be photographed except for the color of the test agent. For this reason, even if the replacement is performed such that the color of the blood is changed to green, there is no concern that the replacement with green does not make it indistinguishable from other green objects. Therefore, in the endoscope image processing, it is possible to perform an extreme hue enlargement process as exemplified in FIG.

  In step S502, the hue adjustment unit 415 expands or reduces the hue by replacing the color using the LUTh, and then uses the hue data H ′ (x, y) converted by the LUTh as an hue component HSV signal. Are converted into RGB signals (S503).

  As LUTh, a plurality of tables designed in accordance with the concept of the doctor who is the user, the design policy of the endoscope manufacturer, the type of test agent used, the site to be observed, the purpose of observation, etc. are prepared. It is preferable. This table may be provided by the manufacturer that provides the electronic endoscope system 1 or may be created by the user himself / herself. The LUTh is stored in the memory 43 of FIG. 2, and is supplied from the microcomputer 42 to the hue adjustment unit 415 in the procedure described with reference to FIG.

  In the above example, the hue adjustment is performed in the HSV space. However, the RGB signal may be converted into the Lab signal, and the a component and the b component may be converted by the lookup table and then returned to the RGB signal.

  Further, a 3D LUT is prepared in the same manner as the standard image generation unit 411 and the color replacement unit 412 without performing conversion to another color space, and the input RGB signal is converted into an RGB signal after hue adjustment. You may make it replace directly.

  As mentioned above, although one Embodiment of this invention was described in detail, the scope of the present invention should be defined by a claim, and is not limited to the said embodiment.

The figure which shows schematic structure of an electronic endoscope system The figure which shows the detailed constitution of the image processing exclusive substrate The flowchart which shows the outline | summary of the initialization process which the microcomputer 32 performs A flowchart showing an outline of processing of the microcomputer 42 The figure for demonstrating the process which converts data into 10 bits The figure which shows an example of the three-dimensional lookup table used in a standard image generation part The figure which shows the other example of the three-dimensional lookup table used with a standard image generation part The figure for explaining the processing method of the value which is not in the lookup table Flow chart showing an outline of processing of the lightness and saturation adjustment unit The figure which shows an example of the look-up table LUTy used by the brightness saturation adjustment part. The figure which shows an example of the look-up table LUTr used in the brightness saturation adjustment part. The figure for demonstrating the relationship between the magnitude | size of an image processing filter, and a processing result Flow chart showing an outline of processing of the sharpness adjustment unit The figure which shows the example of the Gaussian filter used in a sharpness adjustment part Flow chart showing an outline of processing of the hue adjustment unit The figure which shows an example of the lookup table LUTh used by the hue adjustment part The figure which shows the other example of the look-up table LUTh used by the hue adjustment part.

Explanation of symbols

1 Electronic endoscope system 2 Electronic endoscope (scope)
3 Processing unit (processor)
33, 34 selector
4 Image processing board 417a-e Selectors 5, 6 3D lookup table

Claims (3)

An electronic endoscope apparatus having a function of adjusting the hue of an image acquired by an electronic endoscope,
Correspondence between color values that can be obtained by the electronic endoscope by shooting the observation target and color values that are defined so that the hue in the color range necessary for diagnosis is wider than the hue in the color range unnecessary for diagnosis Using the association stored in the replacement table for each inspection purpose in which the attachment is stored and the one replacement table for the inspection purpose specified by the input instruction data, the data is acquired by the electronic endoscope. By providing a purpose-specific image generation means for generating an image for the selected inspection purpose by individually replacing the color value of each pixel constituting the image ,
The electronic endoscope apparatus characterized in that the association of the replacement table is defined so that a hue in a color range unnecessary for diagnosis is reduced .   The electronic endoscope apparatus according to claim 1, wherein the association of the replacement table is defined so that a hue in a color range necessary for diagnosis is expanded. Wherein the association substitution table, the electronic endoscope apparatus according to claim 1 or 2, wherein the defined hue unwanted color range to replace the one predetermined hue diagnosis.

Images