JP2933626B2 - Endoscope image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an endoscope image processing apparatus configured to be hardly affected by noise components.

[Conventional technology]

2. Description of the Related Art Conventionally, as an endoscope image processing apparatus for decomposing an image from an endoscope into a plurality of color signals and converting the signals into electric signals, for example, one having a configuration as shown in FIG. 9 has been proposed.
That is, in FIG. 9, reference numeral 101 denotes a CCD for imaging a living body observation site, which is arranged at the distal end of the endoscope insertion portion. 102 is
An amplifier that amplifies the output signal of the CCD 101, 103 is a γ correction unit, 1
04 is an A / D converter and 105 is a changeover switch. 106,10
7, 108 are R, G, B connected to the output side of the switch 105.
Image memories 109, 110 and 111 of the respective colors are stored in the image memories 106 and 10 respectively.
D / A converter connected to 7,108. 115,116,117
Are contour enhancement circuits, each of which has D / A converters 109, 110, 111
Are connected via changeover switches 112, 113, 114. Reference numeral 118 denotes a control signal generator, and the changeover switches 105, R,
G, B image memory 106, 107, 108, D / A converter 109, 110, 111,
Changeover switches 112, 113, 114, synchronization signal generation circuit 119, and RG
It is connected to a motor 121 that drives the B rotation filter 120. Reference numeral 122 denotes a light source lamp.Light from the lamp 122 is irradiated to the light guide 123 through the filter 120, guided to the end of the endoscope insertion portion, and configured to be illuminated in a frame sequential color system. I have.

In the endoscope image processing apparatus configured as described above,
Illumination light from the lamp 122 is output by the RGB rotation filter 120.
The light is separated into three primary colors of R, G, and B, and sequentially enters the light guide 123 to irradiate the living body. The observation image of the living body is converted into an electric signal by the CCD 101, input to the amplifier 102, is amplified to a voltage level in a predetermined range, enters the γ correction unit 103, and
Will be corrected. The gamma-corrected signal is output from the A / D converter 104.
After the A / D conversion, the signal enters the changeover switch 105, and is sequentially switched by the control signal and is sequentially recorded in each of the R, G, and B image memories 106, 107, and 108. The image signals recorded in the respective R, G, B image memories 106, 107, 108 are sequentially called by the control signal from the control signal generator 118, and the respective D / A converters 109, 110, 111
D / A conversion is performed. The analog image signal is
After passing through the contour enhancement circuits 115, 116, 117 as necessary,
It is sent to the RGB image signal output terminal together with the synchronization signal from the synchronization signal generation circuit 119. The RGB image signal obtained in this way is displayed on a TV monitor for endoscopic observation.

FIG. 10 is a block diagram showing a configuration example of the contour emphasis circuit. The input signal (A) is supplied to the first delay line 201
And the second delay line 202 delays each pixel by one pixel. The output signal (C) from the second delay line 202 delayed by two pixels is the input signal (A)
Is added by an adder 203 to obtain an output signal (D). The output signal (D) is halved by a half inverter 204 and then inverted to obtain an output signal (E). ing. Next, the output signal (E) and the output signal (B) of the first delay line 201 are added by the adder 205 to obtain an outline emphasis component (F). Next, this contour emphasis component (F)
Is amplified to a predetermined size by the multiplier 206 and added to the output signal (B) of the first delay line by the adder 207 to obtain the output signal (G) with the edge emphasized.

Conventionally, not only a method of performing the emphasis processing by hardware in the above-described contour emphasizing circuit but also many methods of performing the contour emphasis processing by software have been proposed. 11 (A) to 11 (C) are explanatory diagrams related to the Laplacian method as an example. That is, FIG.
_ij (where i and j are integers equal to or greater than 1), the target pixel A _ij and its adjacent pixels are _added to the 3 × shown in FIG. 11 (B).
3 is multiplied by each coefficient of the Laplacian matrix. Then, the sum is obtained, and the sum is used as the value of the pixel of interest. By this processing, the outline component can be extracted. FIG. 11 (C) shows an output image processed using the Laplacian matrix.
_{If it is ij} , it is expressed as follows.

_Bij = (0) A _{(i-1) (j + 1)} + (-1) _{Ai (j + 1)} + (0) A _{(i + 1) (j + 1)} + (- 1) A _{(i-1) j} + (4) _Aij + (-1) A _{(i + 1) j} + (0) A _{(i-1) (j-1)} + ( _-1) ) · A _{i (j−1)} + (0) · A _{(i + 1) (j−1)} [Problems to be Solved by the Invention] By the way, in the conventional edge enhancement processing means, attention is paid as described above. A contour component is extracted using a pixel and its adjacent pixels. For this reason, there has been a problem that it is easily affected by the noise component and strongly reacts to a line component or an isolated point finer than the contour component.

In addition, mucosal structures and blood vessel images that are useful information for diagnosis in endoscopic images have a gentle leather-like gradient,
In order to extract such a structural pattern, there has been a problem that the conventional edge enhancement processing method using adjacent pixels has insufficient detection sensitivity.

The present invention has been made in order to solve the above-mentioned problems in the endoscope image processing apparatus using the conventional edge enhancement means, and is less susceptible to noise components, and is useful for medically important mucosal structures and blood vessel images. It is an object of the present invention to provide an endoscope image processing device capable of improving detection sensitivity and performing appropriate image processing even around an imaging region.

[Means and actions for solving the problem]

In order to solve the above problems, the present invention provides an image pickup signal storage unit for storing an output signal of an image pickup unit for picking up an image of a subject, and a method for recognizing a specific point noted from the image signal stored in the image pickup signal storage unit. Pixel data readout means for outputting a readout signal for sequentially reading out pixel data and at least non-adjacent pixel data not adjacent to the target pixel data,
Calculating means for calculating emphasized pixel data obtained by arithmetically emphasizing the specific point based on at least the pixel data of interest and the non-adjacent pixel data read by the pixel data reading means; A peripheral data storage unit that stores peripheral data corresponding to a unit; and a peripheral unit that detects, based on the read signal, that pixel data read from the read unit is peripheral data of an imaging area in the imaging unit. A detecting unit, based on a detection result of the peripheral unit detecting unit, a data selecting unit that selects the emphasized pixel data calculated by the arithmetic unit or the peripheral unit data stored in the peripheral unit data storing unit; The subject image is displayed based on the emphasized pixel data and the peripheral data selected by the data selection unit. And it constitutes an endoscopic image processing apparatus in a video signal generating means for generating a potential of the video signal.

As described above, by extracting the contour component using the information of the pixel that is not adjacent to the target pixel, it is difficult to be affected by the noise component, and a structural pattern having a relatively gentle gradient in a mucous membrane structure, a blood vessel image, or the like is extracted. It becomes possible. Further, since the emphasized pixel data is calculated using the non-adjacent pixel data that is not adjacent to the target pixel data as described above, the non-adjacent pixel data is missing in the periphery of the imaging region, and the correct emphasized pixel data is not obtained. Can no longer be obtained,
According to the present invention, the image processing apparatus further includes means for storing peripheral data corresponding to a peripheral part in the imaging area, and means for detecting that the pixel data to be read is peripheral data of the imaging area. When it is detected, the stored peripheral data is used in place of the emphasized pixel data, so that inappropriate emphasized pixel data is output due to a lack of non-adjacent pixel data in the periphery of the imaging region. Can be prevented.

〔Example〕

Hereinafter, embodiments will be described. FIG. 1 is a schematic diagram showing the overall configuration of one example of a configuration of an endoscope apparatus to which an endoscope image processing apparatus according to the present invention is applied. In the figure, reference numeral 1 denotes an endoscope, its main body 1a is connected to an observation device 2 including an image processing device, and its insertion portion 1b is inserted into a living body 3. The observation device 2 is connected to a TV monitor 4 for observation, and the endoscope main body 1a is connected to an auxiliary device 5 such as a suction device. Then, illumination light is supplied to the end of the endoscope insertion section 1b, and an observation image obtained by irradiating the living body 3 is converted into an electric signal by an image pickup device such as a CCD disposed at the end of the insertion section 1b. , The electric signal is processed in the observation device 2 and converted into a TV signal.
And the endoscope observation is performed.

FIG. 2 is a block diagram showing a first embodiment of the endoscope image processing apparatus according to the present invention. In the figure, reference numeral 11 denotes an A / D converter for converting an analog video input signal into a digital signal. The video signal digitized by the A / D converter 11 is recorded in a main memory 12.
Numeral 13 denotes a read address counter which is a counter for reading a pixel of interest on the main memory 12, and numeral 14 denotes a matrix counter for performing arithmetic processing on the pixel of interest as shown in FIG. , And generates a (9 × 9) size matrix address. Reference numeral 15 denotes a coefficient ROM for storing the coefficients of the (9 × 9) size matrix shown in FIG.
Reference numeral 6 denotes a cumulative multiplier for cumulatively multiplying pixel data in a range of (9 × 9) around each pixel of interest read from the main memory 12 by a coefficient of the coefficient ROM 15. Reference numeral 17 denotes an abnormal value correction circuit for correcting the value of the operation of the accumulative multiplier 16 when the result is out of the predetermined dynamic range. Reference numeral 18 denotes a peripheral part detection circuit that detects a peripheral part of the video data stored in the main memory 12, and 19 denotes a peripheral part data ROM that generates video data of the peripheral part. Reference numeral 20 denotes a data selector, which performs a calculation process on the video data output from the abnormal value correction circuit 17 and the peripheral data output from the peripheral data ROM 19 based on a detection signal of the peripheral detection circuit 18. Switching.
Reference numeral 21 denotes a sub-memory for storing the video data selected by the data selector 20, and reference numeral 22 denotes a D / A converter for converting the video data stored in the sub-memory 21 into an analog signal.

Next, the operation of the endoscope image processing apparatus configured as described above will be described. First, a video input signal for image processing is converted from an analog signal to a digital signal by the A / D converter 11 and recorded in the main memory 12. Image data recorded in the main memory 12, when the address of the pixel of interest is specified by the read address counter 13,
As shown in FIG. 3, the matrix counter 14 centers on the pixel of interest represented by the dot pattern (9 ×
9) By sequentially specifying the addresses of the four pixels,
Reads pixel data from main memory 12 and accumulates multiplier 16
, And the coefficient for each pixel position is read out from the coefficient ROM 15 as shown in FIG.

The accumulative multiplier 16 multiplies (9 × 9) pixel data centered on each target pixel read from the main memory 12 by a coefficient corresponding to each target pixel data read from the coefficient ROM 15 and accumulates the data. Do the addition,
The value is output to the abnormal value correction circuit 17. The abnormal value correction circuit 17 detects whether or not the value calculated by the accumulative multiplier 16 is within the dynamic range at the time of performing display. Correction is made to be a value.

On the other hand, when the target pixel is located in the peripheral portion of the image, correct calculation is not performed because the pixel data around the target pixel is missing. Therefore, the peripheral portion is detected by the peripheral portion detection circuit 18 based on the address value of the read address counter 13, the data for the peripheral portion is read out from the peripheral portion data ROM 19, and input to the data selector 20. Here, the data selector 20 selects the output from the abnormal value correction circuit 17 when the correct arithmetic processing is performed on the pixel of interest. RO
The data from M19 is selected and output to the sub memory 21.
The image data recorded in the sub memory 21 is stored in the D / A converter 2
2 outputs the analog video signal.

Note that the matrix for image processing is not limited to the Laplacian (9 × 9) size matrix shown in FIG. 3, but the conventional matrix shown in FIGS. 4 (A), (B) and (C). Coefficients of various matrices obtained by expanding matrices such as a Laplacian matrix of (3 × 3) size, a first-order differential matrix in the X direction, and a first-order differential matrix in the Y direction are recorded in a coefficient ROM, and several types of matrix counters are provided. By providing the same, similar image processing using various matrices becomes possible. FIG. 5 (A) shows a general-purpose example of the (3 × 3) size matrix shown in FIG. 4 (D), where M × N (M and N are odd numbers of 5 or more and multiples of 3). 3) shows a configuration example of a matrix expanded to a size. FIG. 5 (B) also shows (M−M) between the elements in the X direction of the matrix of the general-purpose example.
3) / 2 zeros and (N-3) between each element in the Y direction
5 shows a configuration example of a matrix expanded by arranging / 2 zeros. In FIG. 5B, M and N are odd numbers of 5 or more.

Further, the matrix used in the arithmetic processing may be configured so that an optimal matrix can be selected according to the angle of view at the time of observation and the type of lesion.

Furthermore, the above-described arithmetic processing need not necessarily be performed on the entire image, and may be configured to be performed only on a required area.

FIG. 6 is a block diagram showing a second embodiment of the present invention, in which the same members as those of the first embodiment shown in FIG. The image processing apparatus according to this embodiment includes an A / D converter 11, a frame memory 31 for recording one frame of a video signal digitized by the A / D converter 11, and a pixel of interest on the frame memory 31. As shown in FIG. 7, a read address counter 13 which is a counter for reading, and 21 lines including (21 × 21) size matrix data centered on the pixel of interest as shown in FIG. A line memory address counter 32 for controlling data capture, and
Line memories 33-1 to 33 for storing data for 21 lines
−21 and (21) around the pixel of interest as shown in FIG.
A matrix horizontal counter 34 for generating an address in the X direction of the (21) matrix, and (21 21) shown in FIG.
And a coefficient ROM 35 for storing each coefficient of the matrix. The other parts are configured in the same manner as in the first embodiment.

Next, the operation of the second embodiment configured as described above will be described. First, a video input signal for image processing is converted from an analog signal to a digital signal by the A / D converter 11 and recorded in the frame memory 31. . Frame memory
When the address of the target pixel is designated by the read address counter 13, the image data recorded in the center 31 is centered on the target pixel by the line memory address counter 32 (21 × 21) as shown in FIG. 21 including pixels on all sides
By sequentially specifying the addresses of the lines, the image data is recorded from the frame memory 31 to the line memories 33-1 to 33-21.

The image data recorded in the line memories 33-1 to 33-21 is read out by the read address counter 13 by sequentially specifying the address of the pixel in the horizontal direction from the matrix horizontal counter 34. The pixel data is read out and input to the accumulative multiplier 16, and the coefficient for each pixel position as shown in FIG. 7 is read out from the coefficient ROM 35 and input to the accumulative multiplier 16. The accumulative multiplier 16 calculates (21 × 21) pixel data centered on the pixel of interest read from the line memories 33-1 to 33-21, and a coefficient corresponding to each pixel of interest read from the coefficient ROM 35. The multiplication and the cumulative addition are performed, and the value is output to the abnormal value correction circuit 17. Subsequent operations are performed in the same manner as in the first embodiment.
Outputs an analog video signal that has been subjected to image processing.

Note that, also in this embodiment, the type of matrix is
The matrix is not limited to the one shown in FIG. 7, and various matrices having expanded sizes of the matrices shown in FIGS. 4 (A), (B) and (c) can be used. In this case, the line memory may be used in accordance with the size of the matrix, and if the matrix size is even larger, it can be added.

In each of the above embodiments, a configuration is described in which one signal is processed as an input signal. However, three such circuits are provided, and each of the R, G, and B signals is processed to perform color display. May be configured. In addition, the present invention can be applied not only to R, G, and B signals but also to complementary color signals such as cyan, magenta, and yellow.
The present invention is also applicable to a signal converted into a chromaticity coordinate system such as a standard. It is also possible to use a composite signal of the NTSC standard or the like for a signal obtained by separating a luminance component and a color component.

In addition, the present invention can perform not only the processing by hardware as shown in the first and second embodiments, but also the processing by software as shown in the flowchart of FIG. 8, for example. Briefly explaining this flowchart, 41
Is a step of selecting the type and size of the matrix used in the conversion processing according to the angle of view and the type of lesion at the time of observation of the input image, 42 is a step of extracting a processing region of the input image, 43 is a step This is a step of determining whether or not the area extracted in 42 is a peripheral part. 44 is a step of performing matrix processing on the extracted region determined to be not the peripheral part in the peripheral part detection step 43 using the matrix selected in the matrix selection step 41, and 45 is a step in which the matrix processing is performed in the entire area. This is a step of determining whether or not the operation has been performed. 46 is a step of detecting the maximum value and the minimum value of the processed pixel values after the matrix processing is completed for all the regions, and 47 is a step of detecting the detected maximum value and the minimum value outside the dynamic range when performing display. Is a step of converting the abnormal value to a value within the dynamic range.

In the processing shown in this flowchart, matrix processing is not performed when a peripheral portion is detected in the processing area.
As in the embodiment shown in the figure, when a peripheral part is detected, data may be read from the peripheral part data ROM to perform image processing.

〔The invention's effect〕

As described above based on the embodiments, the present invention provides a method for at least one of a plurality of video signals forming an image.
Since the configuration is such that the contour component is extracted by using the information of the pixel that is not adjacent to the target pixel, it is hardly affected by the noise component, and it is difficult to react to a fine line component or an isolated point. Further, in a medically important mucosal structure, a blood vessel image, or the like, the detection sensitivity for a structural pattern having a relatively gentle gradient can be increased. Furthermore, even when non-adjacent pixel data is lost in the peripheral portion of the imaging region and correct emphasized pixel data cannot be obtained, the stored peripheral portion data is used instead of the emphasized pixel data. It is possible to prevent output of inappropriate emphasis pixel data due to lack of non-adjacent pixel data in the peripheral portion of the region, and perform appropriate image processing.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing a configuration example of an endoscope apparatus to which an endoscope image processing apparatus according to the present invention is applied, FIG. 2 is a block diagram showing a first embodiment of the present invention, and FIG. FIGS. 4A, 4B, and 4C show the configuration of a (9.times.9) size matrix used for image processing in the first embodiment. FIGS. FIG. 4D is a diagram showing an example of a matrix configuration, FIG. 4D is a diagram showing a general-purpose example of a (3 × 3) size matrix, and FIGS. 5A and 5B are used for image processing in the first embodiment. FIG. 6 shows a configuration example of another matrix, and FIG.
FIG. 13 is a block diagram showing a second embodiment of the present invention, and FIG.
FIG. 9 is a diagram showing a configuration of a (21 × 21) size matrix used for image processing of the second embodiment. FIG. 8 is a diagram showing a flowchart in the case of performing contour component extraction by software. Is a block diagram showing a conventional endoscope image processing apparatus, FIG. 10 is a diagram showing a configuration example of a contour emphasizing circuit in FIG. 9, and FIGS. 11 (A) and (C) are input / output images. FIG. 11 (B) is a diagram showing a configuration of a Laplacian matrix used for processing an input image. In the figure, 1 is an endoscope, 2 is an observation device, 3 is a living body, 4
Indicates a TV monitor, and 5 indicates an attached device.

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-61-121589 (JP, A) JP-A-58-88969 (JP, A) JP-A-60-218171 (JP, A) JP-A 60-218169 218180 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G06F 5/20 H04N 1/40

Claims (1)

(57) [Claims] An image pickup signal storage means for storing an output signal of an image pickup means for picking up an image of a subject, target pixel data of a specific point noted from the image signal stored in the image signal storage means, and at least the target pixel A pixel data readout unit for outputting a readout signal for sequentially reading data and non-adjacent pixel data not adjacent to the data, and the specific point based on at least the target pixel data and the non-adjacent pixel data read out by the pixel data readout unit Calculating means for calculating the emphasized pixel data subjected to the calculation emphasis, peripheral data storage means for storing peripheral data corresponding to a peripheral part in the imaging area of the imaging means, and from the read means based on the read signal. A circuit for detecting that the pixel data to be read is peripheral data of an imaging area in the imaging means. Part detection means, and data selection means for selecting the peripheral data stored in the peripheral data storage means or the emphasized pixel data calculated by the calculation means based on the detection result of the peripheral part detection means, An endoscope image processing apparatus comprising: a video signal generation unit configured to generate a video signal capable of displaying the subject image based on the emphasized pixel data and the peripheral data selected by the data selection unit. .