JPH03121036A - Endoscope image data compressing device - Google Patents

Abstract

PURPOSE:To secure the high degree of compression by suppressing the deterioration of the image quality for a variety of endoscope images by changing the compression methods according to the characteristics of the endoscope image. CONSTITUTION:For example, three compression methods exist, and three kinds of compression rate exist. The compression rate is switched according to the normal image and colored image on the basis of the signal supplied from a compression rate selecting circuit 52. The compression method is selected according to the number of image elements set as one block which is substituted by an average value. For example, if two image elements are set as one block, compression to about 1/2 is performed, and when four image elements are set as one block, compression to about 1/4 is performed, and when nine elements are set as one block, compression to about 1/9 is performed. While, an extension circuit part 36 regenerates the compression discriminating information and the average value in each block from a recording system part 35, and restores the image elements which constitute the block, having the concentration value of each image element in the block as average value, on the basis of the compression discrimination information.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an endoscopic image data compression device that compresses endoscopic image data.

[Prior art and problems to be solved by the invention] In recent years, by inserting an m-long insertion section into a body cavity, VA of internal organs, etc. can be detected, and treatment instruments inserted into treatment instrument channels as necessary have been developed. Endoscopes that can perform various therapeutic procedures are widely used.

Furthermore, electronic endoscopes in which a solid-state carrier water droplet such as a COD is provided at the distal end of the insertion portion have also been put into practical use.

By the way, what about the electron? Endoscopic images taken with an ffti or a TV camera connected to the eyepiece of a fiberscope may be viewed on a TV monitor or recorded on an image recording device for later use in diagnosis and analysis. . When recording endoscopic images in this manner, there is a problem in that a large capacity storage device is required because the amount of image data is large. Furthermore, when transmitting images, there is a problem that the transmission speed is slow.

Therefore, it has been proposed to compress image data.

For example, patent application No. 62-2795 filed earlier by the present applicant.
No. 99 discloses a device as shown in FIG. 38 as a prior art.

In this device, RGB signals constituting an endoscopic image are inputted from an input section 225, converted into digital signals by an A/D converter section 226, and then manually inputted to a compression circuit section 227. The compression circuit unit 227 compresses the image data by predictive encoding or the like, and the compressed image data is recorded in the recording system unit 228. When reproducing an image, the image data on the recording system section 228 is transferred to the decompression circuit section 229.
The image signal is restored to the original image signal by the D/A converter section 231.
The signal is converted into an analog signal and output via the output section 232. Each of the above sections is controlled by a control signal generating section 233. In this device, the A/D converter section 2
R, G in 26.

B The suspension level for each signal is the same.

However, in the case of endoscopic images, there are characteristics such as the R signal is distributed more on the high brightness side and the B signal is distributed more on the low brightness side. If the levels are made the same, there is a problem that some portions of the R and B signals are not effectively used, resulting in poor compression efficiency.

Therefore, the present applicant filed the above-mentioned patent application No. 11f'162-279.
No. 599 proposes an apparatus that performs γ correction and ω conversion according to the characteristics of a plurality of color signals that constitute an endoscopic image.

However, the characteristics of endoscopic images change depending on the observation site, observation method, and the like. In the above device, even though the quantization level is different between R, G, and 8, the quantization level is always unchanged, so it is not always possible to perform optimal compression for various endoscopic images. , the image quality may deteriorate depending on the image.

The present invention has been made in view of the above circumstances, and it is an object of the present invention to provide an endoscopic image data compression device that enables high compression of various endoscopic images with less deterioration in image quality. The purpose is

[Means and effects for solving the problem] The endoscopic image data compression device of the present invention includes a compression means capable of compressing data by changing the compression method according to the characteristics of the endoscopic image. The compression method is changed depending on the characteristics of the data.

[Example] Hereinafter, the present invention will be described in detail with reference to the drawings.

FIGS. 1 to 9 relate to a first embodiment of the present invention.
The figure is a block diagram showing the configuration of the image recording device, FIG. 2 is an explanatory diagram showing the entire endoscopic image filing system, and FIG. 3 is an I! Figure 4 is a block diagram showing the configuration of the image analysis unit, Figure 5 is a histogram of the difference signal q between the normal image and the stained image, and Figure 6 is the recording of the image recording device. 7 is a flowchart showing the reproduction operation of the image recording device, FIG. 8 is an explanatory diagram for explaining the compression operation of the compression circuit, and FIG. 9 is an explanation showing the recording method to the recording system unit. It is a diagram.

As shown in FIG. 2, the endoscopic image filing system includes an electronic endoscope 11, an observation device 3 and a suction device 6 to which the electronic endoscope 1 is connected, and a monitor connected to the observation device 3. 4 and an image recording device 5.

The electronic endoscope vA1 is a IIl device that is inserted into a living body 2, and includes, for example, a flexible insertion section 1a, a large-diameter operation section 1b connected to the rear end of this insertion section 1a, and this operation section. 1b
A connector 1d connected to the 181 detection device 3 is provided at the end of the universal cord 1C.

An illumination window and an observation window are provided at the distal end of the insertion section 1a of the electronic endoscopy vL1. Inside the lighting window,
A light distribution lens (not shown) is attached, and a light guide 18 is connected to the rear end of the light distribution lens. This light guide 18 is inserted through the insertion section 1a, the operating section 1b, and the universal cord 1c, and is connected to the connector 1d. Furthermore, an objective lens system (not shown) is provided inside the observation window, and a solid-state image element, for example, a CCD 8, is disposed at the image forming position of this objective lens system.

The output signal of this CCD 8 is transmitted to the insertion section 1a, the operation section 1b,
The signal is input to the observation device 3 via a signal line inserted through the universal cord 1c and connected to the connector 1d.

The observation device 3 is constructed as shown in FIG.

The observation device 3 includes a lamp 19 that emits white light, and a rotary filter 21 that is provided between the lamp 19 and the incident end of the light guide 18 and is rotationally driven by a motor 20. ing. The rotating filter 21 has red (R) and green (G) filters arranged along the circumferential direction.
), and a filter 22 that transmits light in each wavelength range of blue (B).
R, 22G, 22B, and is rotated by the motor 20 to insert a filter 22R122 into the illumination optical path.
G and 22B are inserted sequentially. The light separated in time series into the RlG and B wavelength regions by the rotating filter 21 is transmitted to the light guide 18. The light is emitted from the distal end of the insertion section 1a of the electronic endoscope 1 after passing through a light distribution lens.

The observation device 3 also includes an amplifier 9, and the output signal of the CCD 8 is amplified to a voltage level within a predetermined range by the amplifier 9, and subjected to γ correction by a γ correction circuit 11. The γ-corrected signal is converted into a digital signal by the A/D converter 12, and then transferred to the memories 14R and 14G corresponding to R, G, and B, respectively, by the changeover switch 13.
, 14B, and the memories 14R, 14G,
14B, Tsuretsure, R image, G image, Bii! 1
IfxI is set to be stored. The memory 1
4R114, G, and 14B are read out simultaneously at the timing of the TV signal, and the D/Δ converter 15.15°15
Each signal is then converted to an analog signal. These analog R, G, and B image signals together with the synchronization signal 5YNC from the synchronization signal generation circuit 16 are
The signal is output from the signal output terminal 17 and input to the monitor 42, image recording device 5, etc. The motor 20
．． A/D converter 12. Changeover switch 13. memory 1
4R114, G, 14B, D/A converter 15. The synchronizing signal generating circuit 16 is controlled by a control signal generating section 23.

Next, the image recording device 5 including the image data compression device will be explained in detail using FIG.

The R, G, and B image signals outputted from the observation device 3 are inputted from the input section 31, and are respectively input to the A/D converter 3.
2, 32, and 32, it is converted into a digital signal and temporarily stored in the R frame memory 33R, the G frame memory 33G, and the B frame memory 338. The R, G, and B image signals read from each frame memory 33R, 33G, and 33B are compressed by a compression circuit section 34, and then recorded in a recording system section 35.

Furthermore, when reproducing image data, the recording system unit 35
The R, G, and B image signals are read out from
The data is decompressed and restored by the decompression circuit section 36. The restored R2G, [3 each image data is
Frame memory 37R for R, frame memory 37G for G
, B frame memory 37B. And this frame memory 37R, 3
7G.

37B, R, G, and B image signals are read out in synchronization with the television signal, and are respectively input to the D/A converter 38.37B.
After being converted into an analog signal at 38.38, it is output from three output sections.

In this embodiment, each of the frame memories 33R133G,
An image analysis section 51 is provided to analyze the characteristics of an endoscopic image from the image information stored in the image information stored in the image information storage device 33B. The output signal of this image analysis section 51 is input to a compression rate switching circuit 52. This compression rate switching circuit 52 is
Based on the signal from the image analysis section 51, the compression ratio in the compression circuit section 34 is determined, and the compression ratio is sent to the compression circuit section 371, and information on the compression ratio of the image is sent to the recording system section 35 as a compression ratio. The recording system section 35 sends this compression rate identification signal as an identification signal to the compressed f'(
, G, and B together with the image information.

Further, a compression rate determination circuit 53 is provided which determines the compression rate from the compression rate identification signal reproduced from the recording system unit 35 and sends information on the compression rate to the decompression circuit unit 36. During playback, the recording system section 35 records compressed R, G, and B data.
A compression ratio identification signal is reproduced together with the image information of the image, and the compression ratio determination circuit 53 determines the compression ratio of the image based on the compression ratio identification signal, and transmits the compression ratio information to the decompression circuit section 3.
Send to 6. This decompression circuit section 36 is configured to perform decompression according to this compression ratio.

Next, the image analysis section 51 will be explained using FIGS. 4 and 5.

As shown in FIG. 4, the image analysis unit 51 includes a 1-pixel delay line 55 that delays the input image signal by one pixel, a subtracter 56 that calculates the difference between the output of the 1-pixel delay line 55 and the input image signal, a comparison circuit 57 that compares the output of the subtracter 56 with a predetermined threshold value;
A counter 58 that counts the output of
8, and a frequency component discrimination signal generation circuit 59 for discriminating frequency components based on the output of the frequency component discrimination signal generation circuit 59, such that the frequency component discrimination signal from the frequency component discrimination signal generation circuit 59 is input to the compression ratio switching circuit 52. It has become.

In this embodiment, the image analysis unit 51 particularly determines whether the endoscopic image is a stained image or a normal image.

In general, a stained image is an image in which the details of the diagnostic site are emphasized compared to a normal image. Therefore, a stained image contains many high frequency components. Therefore, the difference in density values between adjacent pixels is When calculating the histogram of the difference values, in the normal image there are many distributed near O as shown in Figure 5 (a), and in the stained image the values with large absolute values are as shown in Figure 5 (b). Therefore, as shown in Fig. 5(b), a predetermined threshold value is determined, and pixels having a difference in absolute value larger than the threshold value are accumulated. It is possible to distinguish between the two images depending on the magnitude of the values. The image analysis unit 5 shown in FIG. 4 distinguishes between the stained image and the normal image in this way. The difference between both adjacent yarns is determined, and the comparison circuit 57 compares the difference with a threshold value, and the counter 58
Then, the cumulative value of pixels having a difference whose absolute value is larger than the threshold value is determined. Then, the frequency component discrimination signal generation circuit 5
9 outputs a frequency component discrimination signal according to the cumulative value. Incidentally, this image analysis section 51 may analyze all the images of R2O and B, or may analyze only one or two images.
It is also possible to analyze one image.

Next, the operations of the compression circuit section 34 and the expansion circuit section 36 will be explained using FIGS. 6 to 8.

As shown in FIG. 6, the compression circuit section 34 performs step $1
Then, the entire input image is divided into blocks with a predetermined number of pixels, and the average value of the density values of the pixels in each block is calculated. Next, in step S2, the average value is recorded in the recording system unit 35 together with compression identification information based on the compression ratio identification signal from the compression ratio switching circuit 52. In this embodiment, there are three compression methods and three compression ratios. Then, this compression ratio is switched between a normal image and a dyed image based on a signal from the compression ratio switching circuit 52. The compression method is
It is switched depending on whether the fine rain elements are replaced with the average value as one block. For example, if 2 pixels are one block, it will be compressed to about 1/2, if 4 pixels are 1 block, it will be compressed to about 1/4, and if 9 pixels are 1 block, it will be compressed to about 1/9.
Compress it into

On the other hand, as shown in FIG. 5, the decompression circuit section 36 reproduces the compression identification information and the average value of each block from the recording system section 35 in step S3, and in step S4, the decompression circuit section 36 reproduces the compression identification information and the average value of each block based on the compression identification information. The pixels constituting the block are restored using the density value of each pixel as the average value.

FIG. 8 shows an example of compression and expansion with specific density values. (a) Figure relates to the compression method (compression N01) in which 2 pixels are 1 block, (b) Figure is related to the compression method (compression N02) in which 4 pixels are 1 block, and (C) diagram is 9 pixels in 1 block. This relates to a compression method (compression N03) that uses blocks.

(a> As shown in the figure, for compression NO1, Pi, P2
Divide the entire input image into one block, with two pixels of i! 1I Calculate the average value (4) of the IEt values (3, 5), and use this average value (4) as the recording system unit 3.
Recorded in 5. During reproduction, the density value (4) of two pixels is calculated from one average value (4) reproduced from the recording system section 35.
, 4). Similarly, (b) shows t,
Compressed N O2-(-Lt, P n , P 12 ,
P21.

The four pixels of P22 are taken as one block, and the average value (5) of the a degree values (2, 6, 5, 7) of the pixels in the block is recorded in the recording system section 35. During playback, the average value (5) is recorded. From this, the density values (5, 5, 5, 5) of 4 pixels are created. Similarly, (
C) As shown in the figure, in compression NO3, Pu-PI3.
P21～. P23. The nine pixels P3t to P33 are considered as one block, and the density values of the pixels in the block (2, 5, 6
, 6, 4, 7, 4, 3°8) is recorded in the recording system unit 35, and during playback, the average value (5) is 9
Create the density value of the pixel.

The relationship between compression identification information and block size is shown in the table below.

In the case of such compression and expansion, the greater the number of pixels in one block, the higher the compression rate and the worse the resolution during reproduction. The relationship between the number of pixels in one block, the compression ratio, and the resolution during playback for compression Nos. 1, 2, and 3 is as shown in the table below.

During normal imaging, there are few high frequency components, especially the stomach wall, which is a so-called flat image with few high frequency components.
Even if you select compression No. 3, you will hardly notice any deterioration in image quality. Therefore, if the image analysis unit 51 determines that the image is a normal image, compression N03 is selected. On the other hand, with stained images, fine areas become clearer, so compression NO.
If you select 3, the deterioration in image quality will be noticeable. Therefore,
If the image analysis unit 51 determines that the image is a dyed image, the image is compressed N depending on the amount of high frequency components in the image.

Select 1 or 2.

In addition, as shown in FIG. 9, the recording method in the recording system section 35 is to record compression identification information for each image at the beginning, indicating which compression number to use for compression, and then record the average value for each block. shall be recorded. During playback, decompression is performed based on the compression identification information.

In this way, in this embodiment, a normal image and a stained image are automatically distinguished by analyzing the frequency components of an endoscopic image, and the compression method, that is, the compression ratio is changed according to the discrimination result. .

According to the characteristics of the endoscopic image, it is possible to reduce deterioration of image quality and perform high compression suitable for the image.

In addition, in many cases, stained images have a large amount of sunlight component, so R
Alternatively, it may be determined whether the image is a normal image or a stained image based on the size of the B component relative to the G component.

Furthermore, the compression method may be changed between R, G, and B images.

FIG. 10 is a block diagram showing the configuration of an image analysis section in a second embodiment of the present invention.

This embodiment is an example in which the compression ratio is made variable depending on the running state of the blood vessel, and differs from the first embodiment only in the configuration of the image analysis section 51.

Images with many blood vessels running through them have important diagnostic value, so it is necessary to keep the compression ratio low to obtain good image quality. Therefore, in this embodiment, the running state of the blood vessel is automatically determined and the compression ratio is changed.

As shown in FIG. 10, the image analysis section 5 in this embodiment
1 is a differentiation circuit 61 that differentiates an input image signal, a thinning circuit 62 that thins the output image of this differentiation circuit 61,
A binarization circuit 63 that binarizes the output image of this thinning circuit 62, a counter 64 that counts the number of H level pixels in the output image of this binarization circuit 63, and an output of this counter 64. The blood vessel running signal generation circuit 65 generates a blood vessel running signal in accordance with the blood vessel running signal, and the blood vessel running signal is inputted to the compression rate switching circuit 52.

An R image signal containing a large amount of blood vessel information is input to the image analysis section 51, and a differentiation circuit 61 performs differentiation processing on this R image signal to further emphasize blood vessels. Next, a thinning circuit 62 thins the differentially processed image, and a binarization circuit 63 binarizes it. Next, the blood vessel volume is quantified by counting the number of pixels of 1 level in the binarized image using the counter 64.

Then, based on the quantified blood vessel volume, the blood vessel running signal generation circuit 65 generates a blood vessel running signal for changing the compression ratio.

The relationship between the blood vessel i1 compression rate and the resolution during reproduction is as shown in the table below.

Other functions and effects of the configuration 1 are the same as those of the first embodiment.

11 to 14 relate to the third embodiment of the present invention,
FIG. 11 is a block diagram showing the configuration of the compression circuit section;
The figure is a block diagram showing the configuration of the prediction error calculation circuit.
The figure is an explanatory diagram of Iζth illustrating a method of calculating a prediction error, and FIG. 14 is an explanatory diagram of a smoothing filter.

This embodiment differs from the first embodiment in a compression circuit section 34 and an expansion circuit section 36.

As shown in FIG. 11, the compression circuit unit 34 in this embodiment includes a smoothing circuit 41 and a prediction error calculation circuit 42, and the image signals from the frame memories 33R, 33G, and 33B The data is smoothed in 41, predictively encoded in a prediction error calculation circuit 42, and stored in the recording system section 35.

The smoothing circuit 41 has a 3×3 (
Smoothing is performed using a two-dimensional filter (pixel). This filter uses (1-k) times the density value of each pixel as the smoothed density value of each pixel.
Each density value of 8 pixels in the vicinity of that pixel is (k/8
) is the value obtained by adding the multiplied value. Furthermore, k (0<
k<1) is a smoothing coefficient; when this value is large, the smoothing effect is large, and when the value is small, the smoothing effect is small. The value of this smoothing coefficient is switched by a compression ratio switching circuit 52. By arbitrarily setting the value of this smoothing coefficient, the spatial frequency band after smoothing can be determined. That is, the larger k and the greater the smoothing effect, the more the high frequency components of the image deteriorate.

Further, as shown in FIG. 12, the prediction error calculation circuit 42 delays the input data by one pixel by a one-pixel delay line 43, and subtracts this data from the original human data by a reducer 44. It is designed to calculate the difference with the data of the bamboo shoots. As shown in FIG. 13, if the density value of pixel (j) is X(i, j), the prediction error signal ΔX(i, j) output from the prediction error calculation circuit 42 is
j) is Δx (i, j)=x (i, j>-x (i-1, j
). Since this child aIII error signal has a smaller value than the input data, the amount of data recorded in the recording system section 35 can be small.

On the other hand, the decompression circuit section 36 restores the original data by adding the prediction signal line+51 pixels of previous data to the prediction error signal reproduced from the recording system section 35.

Here, if the smoothing coefficient k in the smoothing circuit 41 is increased, the high frequency components of the image will deteriorate, but since the smoothing effect is large, the prediction error signal will become smaller overall.
Therefore, the amount of data to be recorded is reduced. In other words, the compression ratio is high. On the other hand, when k is small and the smoothing effect is small, the high frequency components of the image do not deteriorate, but the prediction error signal becomes large overall, and therefore the amount of data m to be recorded increases. (However, the compression ratio is low. In this way, by arbitrarily setting the smoothing coefficient k in the smoothing circuit 41, the compression ratio can also be set arbitrarily. When there are many high frequency components, the smoothing coefficient k is decreased to lower the compression ratio, and when there are few high frequency components, the smoothing coefficient k is increased to increase the compression ratio.

The other configurations are the same as in the first embodiment.

In this embodiment, the compression ratio is changed depending on the diagnosis site, for example. Generally, when observing the upper gastrointestinal tract, there are many distant images;
When observing the lower gastrointestinal tract, most images are close-up. Therefore, the details of the image obtained when observing the lower gastrointestinal tract are more clearly displayed than those obtained when observing the upper gastrointestinal tract. Therefore, when observing the lower gastrointestinal tract, it is not preferable to increase the compression ratio and deteriorate the image quality.

In this embodiment, when observing the lower gastrointestinal tract, there are many high-frequency components, so this is determined by the image analysis unit 51, and the compression ratio becomes low. On the other hand, when observing the upper gastrointestinal tract, there are fewer high-frequency components, so this is determined by the image analysis unit 51, and the compression ratio becomes higher.

The relationship between the diagnosis site, compression rate, and resolution during playback is as shown in the table below.

Other functions and effects of Structure 2 are the same as those of the first embodiment.

15 to 18 relate to the fourth embodiment of the present invention,
FIG. 15 is a block diagram showing the configuration of the image recording device.
6 is a block diagram showing the configuration of the compression circuit section, FIG. 17 is a block diagram showing the configuration of the band limit switching circuit, and FIG. 18 is an explanatory diagram showing the pass band of each LPF in FIG. 17.

As shown in FIG. 15, in this embodiment, an R band limit switching circuit 67R and a G band limit switching circuit 67G are connected between the input section 31 and the Δ/D converter 32°32.32 in the first embodiment. , B band limit switching circuit 67B is provided. Further, an image signal from the input section 31 is input to the image analysis section 51, and the compression rate switching circuit 52 controls the band limit switching circuits 67R, 67G, and 67B.

Further, the compression circuit unit 34 in this embodiment includes a prediction error calculation circuit 42 similar to that in the third embodiment, as shown in FIG.
However, unlike the third embodiment, there is no smoothing circuit 41. Further, similarly to the third embodiment, the decompression circuit section 36 restores the original data by adding the prediction signal, that is, the data of one pixel before, to the prediction error signal reproduced from the recording system section 35. be.

The band limit switching circuits 67R, 67G, and 67B are configured as shown in FIG. 17.

Each band limit switching circuit 67 (67R, 67G.

Represents 67B. ) is connected to the input end of a one-manpower, two-output changeover switch 70a.

The input ends of a low pass filter (hereinafter referred to as LPF) (1) 68 and an LPF (2) 69 are connected to each output end of the changeover switch 70a, respectively. Each LP
The output terminals of F68 and 69 are respectively connected to respective input terminals of a two-manpower one-output changeover switch 70b. The output of this changeover switch 70b is the output of the band limit changeover circuit 67. The pass bands of each of the LPFs 68 and 69 are as shown in FIG. That is, LPF
(1) 68 removes high frequency components, LPF (2) 69
has a characteristic that it does not remove high frequency components very much.

Further, the image analysis section 51 in this embodiment includes the input section 31
A to convert analog image signals from to digital signals
The configuration is similar to that shown in FIG. 4 or FIG. 12 except that it includes a /D converter, and the frequency components of the image and the running state of blood vessels are determined.

The switches 70a and 70b are the compression ratio switching circuit 52.
It can be switched by.

That is, when the image analysis section 51 determines that the image has few high frequency components or few blood vessels, the switches 70a and 70b select the LPF (1) 68 side, and as a result, the prediction error in the compression circuit section 34 is reduced. The amount of data in the signal becomes smaller. On the other hand, if the image analysis unit 51 determines that the image contains many high-frequency components or contains many blood vessels,
The switches 70a and 70b select the LPF (2>69 side), and as a result, the amount of data B1 of the prediction error signal in the compression circuit section 34 increases, but the image quality does not deteriorate.

In the third embodiment, the band limitation of the image signal is performed by the compression circuit section 34.
However, in this embodiment, the LPF 6 in the band limit switching circuit 67
This is done in an analog manner using 8.69.

Other functions and effects of the configuration 1 are the same as those of the first embodiment.

19 to 23 relate to the fifth embodiment of the present invention,
Figure 19 is a block diagram showing the configuration of the image analysis section;
The figure is an explanatory diagram showing a compression ratio table, Fig. 21 is a flowchart showing the recording operation, Fig. 22 is an explanatory diagram showing the recording method to the recording system section, and Fig. 23 is an explanatory diagram showing the block size. .

In the first to fourth embodiments, the compression rate is made variable for each image unit, but in the fifth to seventh embodiments, the compression rate is made variable for each partial area within the image.

The fifth embodiment is an example in which the compression ratio is made variable between the center and the periphery of the inner 81 mirror image.

This embodiment differs from the first embodiment in the configuration of the image analysis section 51. As shown in FIG. 19, the image analysis section 51 includes a brightness calculation circuit 71 for the central area of the image to which the image signal from the R frame memory 33R is input, and a brightness calculation circuit 72 for the peripheral area of the image. and each calculation circuit 71.
The output of 72 is the flat image/cylindrical image discrimination signal generation circuit 7.
3 is set to be input. The output of this flat image/cylindrical image discrimination signal generation circuit 73 is sent to the compression rate switching circuit 52.

Endoscopic images are roughly divided into two types depending on the observation state. One is an image in which the distance from the tip of the endoscope is approximately the same from the center of the image to the periphery, as in the case of gastric wall observation, and therefore the brightness is approximately constant throughout the image (hereinafter referred to as a flat image).

), and the other is an image (hereinafter referred to as a cylindrical image) that is dark because the center of the image is far away from the tip of the endoscope, and bright because the periphery is close, such as when observing the esophagus. The image analysis section 51 discriminates between the flat image and the cylindrical image.

The recording operation of this embodiment will be explained using FIG. 21.

First, in step 511 (hereinafter, steps will be omitted and simply referred to as S11), the brightness calculation circuit 71 for the central area of the image calculates the brightness of the central area 1 of the image. Let this brightness be A.

Further, at 812, the brightness calculation circuit 72 of the peripheral area of the image
The brightness of the surrounding area of the image is calculated by: Let this brightness be B.

Next, in S13, the flat image/cylindrical image discrimination signal generation circuit 73 determines whether the brightness is A<B, and if YES, it is determined that the image is a cylindrical image, and the information is The data is sent to the compression rate switching circuit 52, and the compression rate switching circuit 52 selects the compression rate table (a> as shown in FIG. The compression ratio switching circuit 52 determines that there is a
This compression rate switching circuit 52 switches to the second compression ratio at 815.
0 Select the compression ratio table (b) as shown in Figure (b).

The compression ratio table is obtained by dividing an image into, for example, 64 parts and determining the compression ratio of each divided image.

The numbers in the figure indicate the compression ratio, and the larger the value, the higher the compression ratio. Therefore, in the compression ratio table (a), the center part is highly compressed and the peripheral part is low compressed.

Furthermore, in the compression rate table (b), the entire image is low compressed.

Next, in S16, the compression circuit unit 34 compresses each divided image according to the compression ratio table selected in 814 or 815.

Then, in S17, the compressed image information is recorded in the recording system section 35 together with the compressed identification information.

In this embodiment, since the compression ratio, that is, the block size, differs for each area within one image, the recording method to the recording system unit 35 is as shown in FIG. 22, for each block.
Compression identification information indicating the compression rate of the block is added before the average value. Also, the block size is
As shown in FIG. 23, three types were used: 1X2.2X2.3X2.

When the image is determined to be a cylindrical image, the relationship among the observed region, compression rate, and resolution at the time of reproduction is as shown in the table below.

The other functions and effects of Structure 2 are the same as those of the first embodiment.

24 to 26 relate to the sixth embodiment of the present invention,
Figure 24 is a block diagram showing the configuration of the image @ analysis section, Figure 25
The figure is an explanatory diagram showing the (RY) (B-Y) plane, and FIG. 26 is a flowchart showing the recording operation.

This embodiment is an example in which the compression ratio is made variable depending on the color.

This embodiment differs from the first embodiment in the configuration of the image analysis section 51. As shown in FIG. 24, the image analysis unit 51 includes each frame memory 33R, 33G, 33B for RGB.
It has a matrix conversion circuit 81 to which RGB image signals from 71 to 81 are input.
The G and B signals are a luminance signal Y and two color difference signals R-Y and B-
It is now converted to Y. This Y, R-Y, B
The −Y signal is recorded in the divided image frame memory 82 and then input to the calculation circuit 83, and the fJ−−Y 2+
(B-Y)2 is now locked out. The ρ calculated by the calculation circuit 83 is inputted to the calculation circuit 84, and ρ is inputted to the calculation circuit 84.
The cumulative value Σg is calculated. The ΣΩ calculated by the calculation circuit 84 is input to a compression ratio determination circuit 85.

In the case of endoscopic diagnosis, color information is extremely important for diagnosis. In other words, information with low saturation such as halation and shadows that is close to black and white does not have much meaning in terms of diagnosis. In particular, the halation area becomes a white area and is meaningless for diagnosis. Therefore, divide the image, calculate the saturation of the image within the divided regions,
If the saturation is low, even if the compression rate is increased and the image quality is reduced somewhat, it will have little effect on diagnosis.

The recording operation of this embodiment will be explained using FIG. 26.

First, in S21, the matrix conversion circuit 81 converts the RG
Convert the B coordinate to Y (RY) (B-Y).

Next, in S22, the divided image frame memory 82 divides the image into, for example, 64 parts.

Next, at 823, the calculation circuit 83 determines that ρ=
−Y)2 + (B−Y)2 is determined.

That is, saturation information is obtained.

Next, in S24, the calculation circuit 84 accumulates g within the divided images.

Next, in 825, the compression rate determination circuit 85 determines the compression rate for each divided image according to the cumulative value Σg of 1.

Next, in 826, the compression circuit unit 34 compresses each divided image according to the determined compression ratio.

Then, in S27, the compressed image information is recorded in the recording system section 35 together with the compressed identification information.

The relationship between saturation, compression rate, and resolution during playback is as follows: The range of compression rate (0, 1, 2, 3> in the above table is
) (B-Y) plane, as shown by the broken line in FIG. 25, for example.

The other functions and effects of configuration 2 are the same as those of the first embodiment.

27 to 30 relate to the seventh embodiment of the present invention,
Figure 27 is a block diagram showing the configuration of the image analysis section.
FIG. 29 is an explanatory diagram showing the (RY) (BY) plane, FIG. 29 is a dimensional explanatory diagram showing the divided image, and FIG. 30 is a flowchart showing the recording operation.

Similar to the sixth embodiment, this embodiment is an example in which the compression ratio is varied depending on the color, but in this embodiment, the compression ratio is set high in areas close to the average color.

This embodiment differs from the first embodiment in the configuration of the image analysis section 51. As shown in FIG. 27, the image analysis unit 51 includes each RGB frame memory 33R, 33G, 33B.
It has a matrix conversion circuit 91 to which RGB image signals from 71 to 91 are input.
The G and B signals are a luminance signal Y and two color difference signals R-Y and B-
It is now converted to Y. This Y, R-Y, B
The Y-fold signal is sent to an average color calculation circuit 92 for all areas and a frame memory 93 for divided images. The output of the divided image frame memory 93 is sent to a divided image average color calculation circuit 94. The average color (X, o, yo
) and the average color (X ij
, yij) are input to the calculation circuit 95, and D = XO−x+J) + (yo −yfj
) 2 is calculated. The calculation circuit 9
g calculated in step 3 is input to the compression ratio determining circuit 96.

In the case of endoscopic diagnosis, the color of the entire image, that is, an area close to the average color, is not very important in the clinic, and an area with a color far from the average color generally indicates a lesion site. Therefore, in a region close to the average color, even if the compression rate is increased and the image quality is slightly degraded, it will hardly affect the diagnosis.

The recording operation of this embodiment will be explained using FIG. 30.

First, at 831, the matrix conversion circuit 91 converts the RG
Convert the B coordinate to Y (RY) (113-Y).

Next, in S32, the calculation circuit 92 calculates (RY)([1
3-Y) Average color of all pixels (Xo) in the plane.

Find yo). Furthermore, (Xo, '10) is the second
As shown in Figure 8, the coordinates of the average color on the (RY) (BY) plane are shown.

Further, in 833, the divided image frame memory 93 divides the image into, for example, 64 parts, and in 834, the average color calculation circuit 94 calculates the average of the pixels in each divided image on the (RY) (B-Y) plane. Color (xij.

yij). As shown in FIG. 29, when the image is divided, the divided image in the i row and j column is Bij, and the coordinates of the average color of the divided image B1j on the (RY) (BY) plane are ( xij, yij).

Next, at 835, the calculation circuit 95 calculates (RY)(B-
Y) Find the distance mρ between the average color of each divided image and the average color of all pixels on the plane.

Next, at 836, the compression rate determination circuit 96 determines the compression rate of each divided image according to the distance ρ.

Next, in 837, the compression circuit unit 34 compresses each divided image according to the determined compression ratio.

Then, in 838, the compressed image information is recorded in the recording system section 35 together with the compressed identification information.

Distance from average color, compression rate, and resolution during playback Also, the range of compression rate in the above table is indicated by the dimensions on the (R-Y) (B-Y) plane and, for example, by the broken line in Fig. 26. It becomes like this.

Other functions and effects of the configuration 1 are the same as those of the first embodiment.

31 to 34 relate to the eighth embodiment of the present invention,
Fig. 31 is a block diagram showing the configuration of the endoscope device;
The figure is a block diagram showing the configuration of the image compression recording unit, FIG. 33 is an explanatory diagram showing the histogram of a general endoscopic image, and FIG.
The figure is an explanatory diagram showing a histogram of a stained endoscopic image.

As shown in FIG. 31, a CCD 101 that converts an image of a living body into an electric signal F3 is provided at the distal end of the insertion section of the endoscope. The output electrical signal of this CCD 101 is input to an amplifier 102 for amplifying it into an electrical signal within a predetermined range (for example, 0 to 1 volt). The output electric signal of this amplifier 102 is connected to the γ correction circuit 103 and the A
After passing through the /D converter 104, the Y signal is input to the selector 105. There are three output terminals of this selector 105, each of which is an R memory 106R.

It is connected to a G memory 106G and a B memory 106B. Each memory 106R, 106G, 106B is connected to D/A converters 107R, 107G, 107B and an image compression recording section 108. The image compression recording section 108 includes an image determination section 121, an image compression section 1222, and an image recording section 123. Said D/A converter 107R
, 107G, and 107B are connected to RGB signal output terminals 109, 110, and 111, respectively.

Further, a control signal generating section 112 is provided which controls the destination of the image signal and the transfer timing when transferring the image signal, and this control signal generating section 112 controls the A/D converter 1o4. Selector 105. RGB each meme memory 106R06G, 1
06B, D/A converter 107R, 107G, 107
B. Connected to the image JIM recording unit 108. The control signal generation section 112 is also connected to a synchronization signal generation circuit 113, and from the synchronization signal generation circuit 113-, a synchronization signal 5YNC for the RGB signal output is outputted to a synchronization signal output terminal 114.

Further, the control signal generation unit 112 includes the RGB rotation filter 1
16 is connected to a motor 115 that drives the motor 16. Light from lamp 118 is passed through RGB rotating filter 116 . The light is emitted from the distal end of the insertion section of the endoscope via the light guide 117 of the endoscope.

Next, the image compression recording section 108 will be explained using FIG. 32.

Each input signal of RG[3 is guided to the peak position detection circuit 140 after passing through histogram creation units 139R, 139G, and 139B, respectively. The output of this peak position detection circuit 140 is transmitted to the selector 132, tt/actor 136 .
It is connected to the II compression information ROM 141. The image determination section 121 is composed of the histogram creation sections 139R, 139G, and 139B, the peak position detection circuit 140, and the compressed information ROM 141. In addition, R memories 131R and 131G for work are provided for each of the six RGB input signals.
After passing through the memory 131G and B memory 131B, it is guided to the selector 132. The output of this selector 132 is the blocking circuit (1) 133. Blocking circuit (2) 134° is connected to blocking circuit (3) 135.

The outputs of the blocking circuits 133, 134, 135 are:
The signal is input to the predictive encoder 137 via the selector 136@. The memory 131R, 1310
, 131B, selector 132° blocking circuit 133,
134,135. Selector 136. Predictive encoder 137
The image compression unit 122 is configured by: and,
The outputs of the predictive encoder 137 and compressed information ROM 141 are recorded in the image recording section 123.

Next, the operation of this embodiment will be explained.

Referring to FIG. 31, the flow of signals will be explained. CCD10
The image signal from 1 is converted by an amplifier 102 into a voltage within a predetermined range, in this example, 0 to 1 volt. This image signal is input to the γ correction circuit 103 and converted into an image signal having a predetermined γ characteristic. After that, in the A/D converter 104, a predetermined quantization level (for example, 8b
digitized by it). Then selector 10
5, the image input to the CCD 101 is stored in the R memory 106R as an image under red (R) illumination, and the image under green (G) illumination is stored in the R memory 106R.
The image with blue (B) illumination is B-106G.
06B, respectively. Each menu 106R,
The signals read from 106G and 106B are sent to the image compression recording unit 108 and D/Δ] inverters 107R, 107G,
107B. This D/A converter 107R
, 107G, RG 13 images from 10713, R
The GB image signal is output from output terminals 109, 110, and 111. On the other hand, a motor control signal is sent from the control signal generating section 112 to the motor 115 that rotationally drives the RGB rotary filter 116.

The motor 115 rotates the RGB rotary filter 116 in accordance with the switching timing of the selector 105 using a control signal. By this RGB rotation filter 116, the illumination light from the lamp 118 is separated into three colors of RlG and B in time series, and guided to the light guide 117 of the endoscope.
It is emitted from the tip of the insertion section of the endoscope. This lighting system is a so-called RGB sequential color system.

Next, the operation of the image compression section 108 will be explained. The signals read from the RGB memory 6R, 106G, and 106B are sent to the working R memory 131R, G memory 131G, and 13-memory 131B in the image compression unit 108 under the control of the control signal generation unit 112. and are recorded in the histogram creation units 139R, 139G, and 139B. In the histogram creation units 139R, 139G, and 139B, RG
B A histogram of each signal is created. Thereafter, the peak position of each histogram is determined by the peak position detection circuit 140, and the selector 132. Selector 136. A control signal is output to the compressed information ROM 141.

On the other hand, signals read from each memory 131R, 131G, and 131B are guided to a selector 132. This selector 132 selects the RGB signals from the blocking circuits (1) 133 . Blocking circuit (2) 134, blocking circuit (3) 13
Lead to one of 5. Three blocking circuits 133
．． 13/l and 135 are each, for example, IX2.2X
Outputs a video signal divided into blocks of 2°3×3 size. The larger the blocking size, the better the compression ratio,
On the contrary, the image quality deteriorates. The selector 136 guides the output of the selected blocking circuit to the predictive encoder 137 based on the control signal of the peak position detection circuit 140. The predictive encoder 137 is configured by [Shokodo Image Processing Handbook No. 217]
A prediction error is obtained using the predictive encoding method described in [Page 219] and the like, and is output to the image recording unit 123. This image recording section 123 is an optical disk,! Data is recorded on a personality recording medium such as an i-disc.

In addition, in order to simultaneously record information such as the block size required for image I and restoration in the image recording unit 123, the peak position detection circuit 140 sends a control signal to the compressed information RoM 141. This compressed information ROMI/11 outputs information such as the blocking size corresponding to the selected output signal to the image recording unit 123.

33(a), (b), and (c) respectively show histograms of RGB components of the -shell endoscopic image;
Figures (a), (b), and (c) each show a histogram of each RGB component of a stained endoscopic image. 33rd
As shown in the figure, in the -film endoscopic image, the R component is biased toward high brightness levels, and the day component is biased toward low brightness areas. Therefore, when the peak position of the RGB three signal histogram is found and the magnitude relationship is examined, R>G>B. On the other hand, when staining with a blue color such as methylene blue is performed, as shown in FIG. 34, the peak positions of His 1 to Gram are approximately equal for B and R, and G is at a low level. However, the magnitude relationship is 8≧R>G.

In this way, from the peak position of the RGB histogram,
- Easily distinguish between membrane endoscopic images and stained endoscopic images.

In general endoscopic images, the R component has few high frequency components, and the B
The component has a low brightness level. Therefore, even if the resolution is lowered for R and 8, it is difficult to visually detect image quality deterioration. Therefore, high compression can be achieved by forming blocks of 2×2 size for the R component and 3×3 size for the B component. On the other hand, the G component has many high frequency components and has a high brightness reflection level. In other words, since image quality deterioration is easily detected visually,
It is possible to compress images with high image quality by forming blocks of ×2 size. Furthermore, in a stained endoscopic image, there are many high frequency components in all three RGB components. Therefore, compression is performed with high image quality by using three components and b1×2 size blocks.

In this way, in this embodiment, depending on the characteristics of the input image,
Select three types of blocking processing.

After that, predictive encoding processing is performed to perform negative layer compression. For this reason, large-sized blocks can be selected for images with high correlation between adjacent pixels and few high-frequency components, such as normal endoscopic images, and high compression can be performed. On the other hand, regarding special images such as those obtained during staining, the correlation between adjacent pixels is low and there are many high frequency components. Therefore, it is possible to select small-sized blocks and perform compression without deteriorating image quality.

From the above, since compression is performed that is suitable for the characteristics of various endoscopic images, it is possible to compress image data with little deterioration in image quality. Furthermore, since three types of compression processing are performed in parallel, the processing time is always constant.

35 to 37 relate to the ninth embodiment of the present invention,
Figure 35 shows the configuration of the image compression recording section (block diagram), Figure 36 is an explanatory diagram showing an endoscopic image in a distant view and its frequency distribution, and Figure 37 shows an endoscopic image in a close view and its frequency distribution. It is an explanatory diagram showing distribution.

This embodiment is the same as the eighth embodiment except that the configuration of the image compression recording section 108 is different.

The configuration of the image compression recording section 108 will be explained using FIG. 35. FFT circuit 159 for each RGB input signal
After passing through R, 159G, and 159B, it is led to a frequency distribution detection circuit 160. The output of this frequency distribution detection circuit 160 is transmitted to the selector 153. Selector 157. The data is input to the compressed information ROM 161. The FFT circuits 159R, 159G, 159
B1 frequency distribution detection circuit 160. Compressed information ROM16
1 constitutes an image determination section 121. Also,
R memory 151 for working with each RGB input signal
R, G memory 151G, B memory 151B, DCT circuits 152R, 152G.

After passing through 152B, it is guided to a selector 153. The output of the selector 153 is sent to the filter circuit (1) 154. The signal is input to filter circuit (2) 155 and filter circuit (3) 156. This filter circuit 15/1. ．． The outputs of 155 and 156 are input to a selector 157. The memories 151R, 151G, 151B, the DCT circuit 152R,
152G, 152B, selector 153. Filter circuit 1
54.155,156. The image compression section 122 is configured by the selector 157. The outputs of the selector 157 and the compressed information ROM 161 are recorded in the image recording section 123.

Third, the operation of the image compression recording section 108 will be explained. The signals read from the RGB meme memories 106R06G and 106B are sent to the working R memories 151R and 151G in the image compression unit 108 under the control of the control signal generation unit 112.
Memory 151G, B memory 151B and FFT circuit 159
R, 159G.

It is recorded in 159B. FFT circuit 159R, 159G
, 159B, Fourier transform is performed on each RGB signal, and its power spectrum is extracted. after that,
The frequency distribution detection circuit 160 determines the frequency distribution range of each signal, and based on this distribution range, the selector 15
3. A control signal is output to the selector 157 and the compressed information ROM 161.

On the other hand, the signals read from each memory 151R, 151G, and 151B are sent to the DCT circuit 152R.

You will be guided to 152G and 152B. Here, for example, rl
EEE Trans Vol. 1G-23, No. 90-9
The 8×8 size discrete COS conversion described in "Page 3" is performed and output to the selector 153. This selector 153 selects the RGB signal from the filter circuit (1) 154 based on the control signal of the frequency distribution detection circuit 160.
．． Filter circuit (2) 155. Filter circuit (3) 15
Lead to one of 6. The three filter circuits 154°, 155, and 156 are arranged, for example, in a 2×2.
It is a transmission filter with a size of 3x3.4x4. The smaller the filter size, the better the compression rate, but the lower the image quality. The selector 157 is a frequency distribution detection circuit 160
Based on the control signal, the output of the selected filter circuit is output to the image recording section 123. On the other hand, the frequency distribution detection circuit 160 sends a control signal to the compressed information ROM 161 in order to simultaneously record information such as the filter size required at the time of restoration. The compressed information ROM 161 outputs information such as a filter size corresponding to the selected output signal to the image recording unit 123.

Here, the frequency distribution of endoscopic images will be explained using FIGS. 36 and 37. In this example, consider a case where the same subject is observed at different observation distances. 36th
Figures (a) and (b) respectively show an endoscopic image in a distant view and a power spectrum showing its frequency distribution.
Figure 7(a).

(b) shows an endoscopic image in a close view and a power spectrum showing its frequency distribution, respectively. During a distant view, high frequency components such as the mucous membrane structure of the living body are remasked by the resolution of the optical system and are not detected. When the frequency distribution in this case is imaged as a power spectrum, the image is concentrated at the origin, that is, at low frequency components, as shown in FIG. 36(b). On the other hand, in close view, high frequency components such as the mucous membrane structure of the living body are detected. The frequency distribution in this case is distributed over a wide range centered on the origin, as shown in FIG. 37(b). By obtaining the power spectrum in this way, the proportion of high frequency components can be determined.

In endoscopic images, even when photographing the same subject, the amount of information contained in the video signal differs depending on the observation distance. That is, when the observation distance is close and there are many high frequency components, compression with high image quality can be achieved by setting the filter size to 4×4. Conversely, when the observation distance is long and there are few high frequency components, high compression can be achieved by setting the filter size to 2×2.

Furthermore, even at the same observation distance, there is a large difference in the amount of information between the upper and lower digestive tracts. This is because blood vessel images are hardly detected in the upper gastrointestinal tract such as the stomach, but blood vessel images are detected in the lower gastrointestinal tract such as the large intestine. In the lower gastrointestinal tract where blood vessel images are detected, there are many ^ frequency components, and in the upper gastrointestinal tract where blood vessel images are not detected, there are few high frequency components. This allows the filter size to be increased, for example, by 2× in the upper gastrointestinal tract
2. In the lower gastrointestinal tract, the image size is 3×3, making it possible to balance image quality and compression rate.

The other operations and effects of the configuration 9 are the same as those of the eighth embodiment.

Note that the present invention is not limited to a frame-sequential electronic endoscope using RGB signals, but can also be applied to a single-panel electronic endoscope that decodes a composite video signal. Further, the endoscope may be of a type having an image pickup device at its tip, or of a type having an image guided to the outside of the object to be observed via an image guide using an optical fiber and then received by the image pickup device.

Effects of E-firing] As explained above, according to the present invention, the compression method can be changed depending on the characteristics of the endoscopic image, thereby reducing deterioration in image quality for various endoscopic images. This has the effect of enabling high compression.

[Brief explanation of drawings]

FIGS. 1 to 9 relate to a first embodiment of the present invention.
The figure is a block diagram showing the configuration of the image recording device. An explanatory diagram showing the entire one-image filling system, Fig. 3 is a block diagram showing the configuration of the observation device, Fig. 4 is a block diagram showing the configuration of the image analysis section, and Fig. 5 shows the normal image and stained image. A histogram of the difference signal, FIG. 6 is a flowchart showing the recording operation of the image recording device, FIG. 7 is a flowchart showing the reproduction operation of the image recording device, FIG. 8 is an explanatory diagram for explaining the compression operation of the compression circuit, FIG. 9 is an explanatory diagram showing the recording method to the recording system section, FIG. 10 is a block diagram showing the configuration of the image analysis section in the second embodiment of the present invention, and FIGS. Regarding the third embodiment, FIG. 11 is a block diagram showing the configuration of the compression circuit section, FIG. 12 is a block diagram showing the configuration of the prediction error calculation circuit, and FIG. 13 is an explanation for explaining the prediction error calculation method. 14 is an explanatory diagram of a smoothing filter, FIGS. 15 to 18 relate to the fourth embodiment of the present invention, FIG. 15 is a block diagram showing the configuration of an image recording device, and FIG. 16 is a compression circuit. 17 is a block diagram showing the configuration of the band limit switching circuit, and FIG. 18 is a block diagram showing the configuration of the band limit switching circuit.
19 to 23 relate to the fifth embodiment of the present invention, FIG. 19 is a block diagram showing the configuration of the image analysis section, and FIG. 20 is a compression ratio. FIG. 21 is an explanatory diagram showing the table, FIG. 21 is a flowchart showing the recording operation, FIG. 22 is an explanatory diagram without showing the recording method to the recording system part, FIG. 23 is an explanatory diagram showing the block size,
4 to 26 relate to the sixth embodiment of the present invention, and the second embodiment
4 is a block diagram showing the configuration of the image analysis section, FIG. 25 is an explanatory diagram showing the (RY) (B-Y) plane, FIG. 26 is a flowchart showing the recording operation, and FIGS. 27 to 30 The figures relate to the seventh embodiment of the present invention, FIG. 27 is a block diagram showing the configuration of the image analysis section, and FIG. 28 is (R-Y) (B-Y
) An explanatory diagram showing a plane, Fig. 29 is an explanatory diagram showing a divided image, Fig. 30 is a flow diagram showing a recording operation, and Fig. 31 is an explanatory diagram showing a divided image.
The figures to 34 relate to the 8th embodiment of the present invention, and the 31st embodiment
Figure 32 is a block diagram showing the configuration of the endoscope device, Figure 32 is a block diagram showing the configuration of the image compression recording section, Figure 33 is a histogram of a general endoscopic image, and Figure 34 is an explanatory diagram. is an explanatory diagram showing a histogram of a stained endoscopic image, No. 35
Figures 37 to 37 relate to the ninth embodiment of the present invention;
The figure is a block diagram showing the configuration of the image compression recording unit, and FIG. 36 is an explanatory diagram showing an endoscopic image in a distant view and its frequency distribution.
FIG. 37 is an explanatory diagram showing an endoscopic image in a close view and its frequency distribution, and FIG. 38 is a block diagram showing a conventional double-image compression device. 1... Inside the electron? Ji mirror 5...image recording device 3
4... Compression circuit section 35... Recording system section 51
...Image analysis section 52...Compression rate switching circuit 2nd
Figure 3 Figure 6 Figure 7 Figure 4 (a) (b) 255 difference value +6 -255 Shihi turtle 0 1/deiyu +255 difference value Figure 8 (0) (b) (C ) Figure 11 Figure 13 Figure 14 Figure 17 Figure 8 Old figure Figure 21 Figure 22 1 1,000 households

Claims (1)

[Claims] An endoscopic image data compression device for compressing image data of endoscopic images obtained by an endoscope, which is equipped with a compression means capable of compressing data by changing compression methods according to the characteristics of the endoscopic image. An endoscopic image data compression device characterized by: