JPH10232929A - Method and device for processing image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prepare blurred images at a high speed in the case of preparing the blurred images of a large blur amount by performing a blurring processing by reduced images obtained by reducing the number pixels of input image and turning the blurred reduced images to an original number of pixels. SOLUTION: Images are inputted in an image input process P0, a reduction rate is obtained from a blur amount inputted in a blur amount input process P1 and the blur amount is decided based on the blur amount from the input process P1 and the reduction rate from a reduction rate decision process P2 in a reduced image blur amount decision process P. Then, the input images are reduced based on the reduction rate decided in the image reduction rate decision process P2 in an image reduction process P4 and the reduced blurred images are prepared in a reduced image blurring process P5 from the reduced images reduced in the image reduction process P4 by the reduced image blur amount. Then, the reduced blurred images prepared in the reduced image blurring process P5 are magnified to the same size as the input images in an image magnification process P6 and magnified blurred images are outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the formation of an unsharp signal corresponding to a blurred image and an image processing method and apparatus using the unsharp image.

[0002]

2. Description of the Related Art Conventionally, as an image enhancement processing method,
The unsharp masking process is well known and is often used in photographic technology and the like. Recently, it has also been used as a method of enhancing digital images in a medical X-ray imaging system (edited by Takagi, Toriwaki and Tamura) "Latest trends in image processing algorithms": New Technology Communications).

Now, assuming that an input image is f (x, y),
Processing result image g in unsharp masking processing
(X, y) is represented by the following equation (1).

G (x, y) = f (x, y) + c × {f (x, y) −f _av (x, y)} (1)

Here, f _av (x, y) is a local average value at the point (x, y) and is obtained from an n × m pixel area around the point (x, y). It can be calculated using a simple average pixel value as shown in equation (2).

F _av (x, y) = {1 / (n × m)} × ΣΣf (xi, yi) (2)

The local average value f _av (x, y) indicates a blurred image obtained by blurring the input image f (x, y). The larger the peripheral pixel area for obtaining the local average is, the more blurred the image is obtained. Can be Further, the second term in the equation (1) includes a high-frequency component of the input image due to the difference, and the unsharp masking processing is performed by adding a high-frequency component multiplied by a coefficient c to the input image.

[0008] Further, in a medical X-ray imaging system, an imaging method using an analog filter to better observe the mediastinum in chest simple imaging and chest tomography as well as unsharp masking processing is known. This is known, and can be realized by performing image processing on an image acquired by normal photographing without using a filter. (Otani, others "C
R, Development of Self-Compensating Digital Filter ”, Journal of the Radiological Technology Society of Japan, Vol. 45, No. 8, p. 1030,
1989)

The processing result image g '(x, y) in this self-compensation filter processing is represented by the following equation (3), where f (x, y) is an input image.

G ′ (x, y) = f (x, y) tens F {f _av (x, y)} (3)

Here, f _av (x, y) is a local average value at the point (x, y) as in the equation (1).
(X, y) represents a blurred image. Also, F
{*} Is a function meaning an analog filter.

[0012]

However, the unsharp masking process and the self-compensation filter process in the above-described conventional example are blurred image processes represented by f _av (x, y) in each case, and therefore, a blurred image is created. The amount of calculation is the dominant factor in the processing time.

Further, when a blurred image is created, if the pixel area n × m for obtaining the local average value at the time of creating the blurred image is increased in order to obtain an image having a larger blur amount, the processing speed is extremely slow. There is a problem. For example, if the image area n × m for obtaining the local average is k times (an integer of k> 0) both vertically and horizontally, the area for obtaining the local average is kn × km
= K ² (n × m), and there is a problem that the calculation amount for obtaining the local average value at a certain point (x, y) is k ² times.

An object of the present invention is to solve the above-mentioned problems,
It is an object of the present invention to provide a method and apparatus for creating a blurred image at high speed even when creating a blurred image having a large blur amount.

[0015]

According to the present invention, there is provided an image processing method comprising the steps of: inputting an image signal; and reducing the number of pixels in the input image signal. A reduced image blurring step of performing a blurring process on the image signal output by the image reduction step, and an image enlarging step of increasing the number of pixels of the image signal output by the reduced image blurring step and outputting a blurred image signal And
It is characterized in that a blurred image signal blurred with a desired blur amount is obtained.

According to another aspect of the present invention, there is provided an image processing apparatus comprising: input means for inputting an image signal; reducing means for reducing the number of pixels of the input image data; and blur processing for the image data output by the reducing means. , And increasing means for increasing the number of pixels of the image data output by the blur processing means, to obtain a blurred image signal blurred by a desired blur amount.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail based on the illustrated embodiment. FIG. 1 is a block diagram showing a configuration of an entire image processing apparatus as one embodiment of the present invention.
Reference numeral 1 denotes a central processing unit (CPU) that performs overall control of image processing, and a work storage device 2 that functions as a main storage area or a work area of the central processing device 1, for example, an image such as a scanner or a camera. An image input device 3 for input, an image output device 4 for outputting a processing result to a printer or a film, an operation input device 5 for inputting a blur amount to the device with a keyboard, a mouse, or the like, or a device such as a CRT or a liquid crystal monitor. Each output of the operation content display device 6 for displaying a value to be input to the device, a state of the device, and the like is output from the system bus 7.
Are connected to each other. In the apparatus according to the present embodiment, image data input from the image input device 3 or image data stored in the external storage device 8 can be processed as input image data.

FIG. 2 is a flowchart for explaining the operation of the apparatus shown in FIG. 1, and performs processing on a predetermined input image S0 input in the image input process P0. First, a desired blur amount S1 is input in a blur amount input step P1, a reduction ratio P2 for reducing an image in a reduction ratio determination step P2 is obtained from the blur amount S1, and a blur amount S1 and a reduction from the input step P1 are obtained. In the reduced image blur amount determining step P3, a blur amount S3 for performing a blurring process on the reduced image is determined based on the reduction rate S2 from the rate determining step P2. Note that the reduction process described here is a process of reducing the number of pixels to be used in subsequent processes as described later.

Next, in the image reduction process P4, the input image S0 is determined based on the reduction ratio S2 determined in the image reduction ratio determination process P2.
Is reduced to S4, and a reduced blurred image S5 is created in a reduced image blurring step P5 from the reduced image blurring amount S3 determined in the reduced image blurring amount determining step P3 and the reduced image S4 reduced in the image reducing step P4. Then, the reduced blurred image S5 created in the reduced image blurring step P5 is enlarged to the same size as the input image S0 in the image enlargement step P6, and the enlarged blurred image S6 is output. Note that the enlargement process is a process of increasing the number of pixels to be equal to the number of pixels of the input image, as described later.

Now, the blur amount S1 input in the blur amount input step P1, that is, the size of the pixel area N × for obtaining the local average value
Let M be a 15 × 15 pixel area and let the size of the input image be X ×
Assuming that Y (X> N, Y> M), the reduction ratio h for reducing the input image to 1 / h is an integer that does not exceed the desired blur amounts N and M and is an integer that can divide N and M. Just choose. That is, since N = M = 15, if N and M are dividends, an integer that is divisible and does not exceed N and M is selected. Here, for example, the reduction ratio h = 5
Shall be selected.

Next, in a reduced image blur amount determining step P3, a blur amount S3 to be applied to the reduced image is determined.
That is, from the reduction ratio h determined in the reduction ratio determination step P2, the reduced image blur amount S3 may be determined to be (N / h) × (M / h), and since 15/5 = 3, the reduced image blur amount S3 is 3x3
Becomes

Next, in an image reduction step P4, a reduced image of the input image is created by an averaging method based on the reduction rate h determined in the reduction rate determination step P2. That is, the reduction ratio h =
Since it is 5, the pixel values of the 5 × 5 pixel area of the input image are averaged to obtain the pixel values of the reduced image. FIG. 3 shows that the average value of all pixel values (25 pixels) in a 5 × 5 pixel area (area separated by a thick line) is used as the pixel value of the reduced image.

Next, in a reduced image blurring step P5, the reduced image of the input image data S0 is blurred by a predetermined blurring method, for example, a method as shown in equation (2). Thereafter, in an image enlargement step P6, the blurred image S5 of the reduced image created in the reduced image blurring step P5 is magnified five times from the reduction rate h determined in the reduction rate determination step P2. An example of the enlargement method is (Introduction to Computer Image Processing, supervised by Tamura: Soken Publishing)
Use a linear interpolation method as described in.

FIG. 4 is an explanatory diagram of the linear interpolation method, and shows a 3 × 3
Enlarge the pixel area using linear interpolation to get
Consider making a pixel.

The coordinates (i, j), (i + 1, j), (i, j)
(i + 1), (i + 1, j + 1)
(s, i + t): When the pixel value of 0 ≦ s, t ≦ 1 is obtained by the linear interpolation method, it is expressed as the following equation (4).

F (i + s, j + t) = f (i, j) × (1-s) × (1-t) + f (i + 1, j) × s × (1-t) + f (i, j + 1) × ( 1−s) × t + f (i + 1, j + 1) × s × t (4)

As an image enlargement method, there are various enlargement methods such as a zero-order interpolation method and a tertiary interpolation method other than the linear interpolation method.
In consideration of the purpose of creating a blurred image, a method other than the linear interpolation method is not appropriate because the processing itself has a characteristic of generating a high-frequency component.

Next, a calculation is made as to how much calculation amount can be reduced. First, C _org can be approximately calculated by the following equation (5) when the input image is directly calculated with the blur amount of 15 × 15 pixels. This equation (5) means that the number of calculations MN for creating one pixel of a blurred image is processed for XY pixels.

C _org = MN × XY = 15 ² XY = 225XY (5)

Then, the amount of calculation C for creating the reduced image
_min is expressed by the following equation (6).

[0031] ^{_{C min = h 2 × (XY}} / h 2) + (NM / h 2) × (XY / h 2) + 16h 2 × (XY / h 2) = 17XY + (NMXY / h 4) = 17XY + (9XY /25)=17.4XY (6)

[0032] Formula first term of (6), in order to obtain one pixel of the reduced image, by performing a calculation from the reduction ratio h h ² times, and calculates only the total number of pixels XY / h ² min of the reduced image It represents that. The second term represents the number of calculations for the blurring process on the reduced image. In one pixel of the reduced image, (N /
h) × (M / h) indicates that the average value in the area is obtained, and the average value is calculated for (X / h) × (Y / h) pixels. The third term is a reduced image (X / h) × (Y /
h) represents the number of operations when expanding to X × Y pixels. In order to obtain one pixel by the linear interpolation method, a pixel value to be interpolated from four surrounding pixels is obtained based on Expression (4), and h ²
The calculation is performed for (X / h) × (Y / h) pixels. Thus, it can be seen from equations (5) and (6) that the amount of calculation can be reduced by about 1/12.

As described above, in order to create a blurred image,
A reduced image of the input image is created, a blur process is performed on the reduced image, and the reduced blurred image is enlarged to create a desired blurred image. Thus, it becomes possible to obtain a blurred image equivalent to a case where the input image is directly subjected to the blurring processing.

The blurred image data created in this way is used as an unsharp signal in the image correcting step P7. Here, in the image correction step P7, the following image correction processing is performed using the image data linearly interpolated in the image enlargement step.

For example, as a first example, it is possible to enhance the edge (high-frequency portion) in order to adjust the image correction processing. That is, the image input step P0
Let f (x, y) be the image data input in step (1) and f _av (x, y) be the unsharp signal.
As shown in (1), the processed image g (x, y) is converted to f (x, y).
y) + k × {f (x, y) -f _av (x, y)} (k is a coefficient), and a so-called unsharp masking process is used to emphasize high-frequency components in a required portion.

In the next example, the unsharp signal is subjected to a filtering process as shown in equation (3) using f _av (x, y) to increase the contrast of the image of a specific portion. It is also possible.

Further, as a next example, it is possible to compare the unsharp signal with a separately determined threshold value and use the area division of the photographed image. For example, a histogram of an input image signal is taken for separation and identification of a lung field portion and a mediastinal portion, a value corresponding to a valley of the histogram is calculated, and the calculated value is used as a threshold to binarize an unsharp signal. , Separates the image into two regions. Then, by applying different processing to each of the two separated regions, the entire image is corrected to an image suitable for diagnosis.

In the above embodiment, the image reduction step P4
Although the averaging method was used to reduce the input image of the above, in the case of the input image from which the high-frequency components have been sufficiently removed, even if the input image is simply thinned out, it is necessary to create a blurred image. Is not a problem.

FIG. 5 is an explanatory diagram of a method of simply thinning out an input image. Assuming that a certain pixel in a 5 × 5 pixel area (an area surrounded by a thick line) is a pixel value of a reduced image, the following equation (6) is used. The indicated number of operations C _min is as shown in the following equation (7).

C _min = (XY / h ² ) + (NM / h ² ) × (XY / h ² ) +16 h ² × (XY / h ² ) = (XY / h ² ) × {(1+ (NM / h ² )｝ + 16XY = (XY / 25) × (1 + 9) + 16XY = 16.4XY (7)

If this method is used, the amount of calculation can be reduced by about 1/13 as compared with the equation (5), and the calculation speed is improved.

In the above-described embodiment, the image processing is performed by the C processing shown in FIG.
Although processing is performed by software mainly using a PU, the idea of the present invention can be similarly applied to an apparatus that performs image processing by hardware.

FIG. 6 is a block diagram showing an image processing apparatus as another embodiment. In the figure, reference numeral 11 denotes an input terminal of a digital image signal to be processed, and reference numeral 12 denotes the digital image signal composed of, for example, 5 × 5 pixels. It is a blocking circuit for blocking into blocks, and this blocking circuit 12
A buffer memory capable of storing at least 5 lines of image data. In this buffer memory, the order is changed so that input image data input in raster order is sequentially output in blocks of 5 × 5 pixels. Circuit.

The output of the blocking circuit 12 is input to an average value calculation circuit 13, which outputs 2
The average value is calculated for every five pixels, that is, for each block. Therefore, the number of samples output from the average value calculation circuit 13 is 1/25 of the number of input pixels. Reference numeral 14 denotes a blurring circuit to which the average value output from the average value calculating circuit 13 is input, and the image in which the number of samples is reduced is further filtered to blur the entire image.

The number of samples output from the blur processing circuit 14 is naturally 1/25 of the number of samples of the input image.
Even if a digital filter having a considerably complicated configuration is used as the blur processing circuit 14, high-speed processing is possible, and the circuit scale is greatly reduced. Next, the output of the blurring processing circuit 14 is stored in the memory 15. This memory 1
In writing to 5, the writing is executed while skipping the pixel to be interpolated as the address, and the linear interpolation circuit 16 accesses the pixel data written in the memory 15 by the method described above. The interpolation pixel is linearly operated, and the interpolation pixel is written to the remaining address.

As described above, the unsharp signal stored for one screen in the memory 15 is read out in synchronization with the input image signal stored for one screen in the memory 17 and input to the image quality adjustment processing circuit 18. Is done. The image quality adjustment processing circuit 18 performs image correction processing according to the above-described equations (1) and (3), and performs different processing on an input image signal depending on a signal obtained by binarizing an unsharp signal. The output of the image quality adjustment processing circuit 18 is supplied from an output terminal 19 to an image output device such as a printer or a display.

Also in the embodiment constituted by the hardware as described above, the circuit scale of the blur processing circuit 14 can be reduced, and the processing speed of the entire apparatus can be improved by increasing the processing speed.

[0048]

As described above, the image processing method and apparatus according to the present invention perform blur processing on a reduced image obtained by temporarily reducing the number of pixels of an input image, and reduce the blurred reduced image to the original number of pixels. By doing so, it is possible to create a blurred image equivalent to the case where the input image is directly blurred, it is possible to perform it at a high speed with a smaller number of calculations, for example, a blurred image such as unsharp masking processing or self-compensation filter processing Can be performed more efficiently.

[Brief description of the drawings]

FIG. 1 is a block diagram of the overall configuration of an image processing apparatus according to an embodiment.

FIG. 2 is a flowchart for explaining the operation.

FIG. 3 is a diagram illustrating an example of a process of reducing the number of pixels by extracting an average pixel, which is performed in an image reduction process.

FIG. 4 is an explanatory diagram of a linear interpolation method performed in an image enlargement step.

FIG. 5 is an explanatory diagram of another example of a pixel number reduction process by extracting an average pixel, which is performed in the image reduction process in FIG. 2;

FIG. 6 is a block diagram of the overall configuration of an image processing apparatus according to another embodiment.

[Description of Signs] 1 CPU (central processing unit) 2 Work storage device 3 Image input device 4 Image output device 5 Operation input device 6 Display device 7 System bus 8 External storage device 12 Blocking circuit 13 Average value calculation circuit 14 Blur Processing circuit 16 Linear interpolation circuit 18 Image quality adjustment processing circuit

Claims (18)

[Claims] 1. An image signal inputting step of inputting an image signal, an image reducing step of reducing the number of pixels of the input image signal, and a reduced image blur performing a blurring process on the image signal output by the image reducing step. And an image enlargement step of outputting a blurred image signal by increasing the number of pixels of the image signal output by the reduced image blurring step, to obtain a blurred image signal blurred with a desired blur amount. Image processing method. 2. The image processing method according to claim 1, wherein in the image enlarging step, the number of pixels of the image signal is increased by performing linear interpolation using the reduced image signal blurred in the reduced image blurring step. . 3. The image processing according to claim 1, wherein in the image reduction step, an average value of a plurality of pixels in the input image signal is calculated, and the average value is output instead of the plurality of pixels. Method. 4. The image reduction process according to claim 1, wherein only one pixel is extracted from a plurality of pixels in the input image signal, and only the extracted pixels are output instead of the plurality of pixels. Image processing method. 5. An image correction step for correcting an image signal input in an image signal input step using an unsharp signal output in the image enlargement step.
The image processing method according to claim 1. 6. The image processing method according to claim 5, wherein in the image correcting step, a process of enhancing a high-frequency component of the input image signal is performed. 7. The image processing method according to claim 5, wherein in the image correction step, a process of adjusting a contrast of the input image signal is performed. 8. The image processing method according to claim 5, wherein in the image correction step, a process of distinguishing the input image signal into a plurality of regions is performed. 9. An input unit for inputting an image signal, a reducing unit for reducing the number of pixels of input image data, a blur processing unit for performing a blurring process on image data output by the reducing unit, and the blurring unit An image processing apparatus comprising: an increasing unit that increases the number of pixels of image data output by the processing unit, and obtains a blurred image signal with a desired blur amount. 10. The image processing apparatus according to claim 9, wherein the increasing unit increases the number of pixels of the image data by performing linear interpolation using the image data output by the blurring unit. 11. The image processing apparatus according to claim 9, wherein the reduction unit includes an average value unit that calculates an average value of a plurality of pixels in the input image data, and outputs the average value instead of the plurality of pixels. The image processing apparatus according to any one of the preceding claims. 12. The image processing apparatus according to claim 9, wherein said reducing means includes extracting means for extracting only one pixel from a plurality of pixels in the input image data, and outputs only extracted pixels instead of said plurality of pixels. An image processing apparatus according to claim 1. 13. The image processing apparatus according to claim 9, further comprising a correction unit configured to correct the image data input by the input unit using the image data output by the increase unit. 14. The image processing apparatus according to claim 13, wherein the correction unit performs a process of enhancing a high frequency component of the input image data. 15. The image processing apparatus according to claim 13, wherein the correction unit performs a process of adjusting a contrast of the input image data. 16. The image processing apparatus according to claim 13, wherein the correction unit performs a process of distinguishing the input image data into a plurality of areas. 17. The image processing apparatus according to claim 9, wherein said blur processing means is constituted by software. 18. The blur processing means is constituted by a digital filter using hardware.
The image processing device according to claim 1.