KR100416254B1 - Appatatus for detecting high frequency elements from image data and method thereof - Google Patents

Abstract

An apparatus and method for detecting high frequency components of image data are disclosed. The window setting unit sets the size and position of the window represented by N × N. The first calculation unit calculates an average value of pixel values of pixels existing in the N × N window. The second calculation section calculates an allowable range value for pixel values of pixels existing in the N × N window. The detector detects a pixel having a high frequency component based on a frequency of pixels deviating from an allowable range value input from the second operator, based on the average value input from the first operator. According to the present invention, by accurately detecting the high frequency component corresponding to the boundary of the image by using the average value and statistical characteristics of the surrounding pixels of the pixel to be irradiated, it is possible to save the sharper image by utilizing the boundary of the image. Furthermore, the text component can be effectively expressed by preserving the text component detected at the time of scaling the image.

Description

Apparatus for detecting high frequency elements from image data and method

The present invention relates to an apparatus and method for detecting high frequency components of image data, and more particularly, to an apparatus and method for detecting boundary regions and character regions that are high frequency components of input image data.

In order to reduce storage capacity and transmission efficiency in a general video system, the original video is compressed and stored or transmitted. At this time, the same image as the original image should be output when reproducing the stored image data or restoring the received image data. In particular, in the transmission of image data, there is a need to reduce the influence of noise.

In such image restoration, factors affecting the quality of the reconstructed image include color, color tone, image blur, and sharpness. Among them, whether the boundary of the image is reproduced is an important problem with regard to the image clarity. In general, the boundary portion of the image data may be distinguished by a high frequency component. That is, the pixel of the high frequency component is much larger or smaller than the surrounding pixels. Therefore, if the high frequency component can be accurately separated from the image data, it is possible to provide a clearer image.

1 is a block diagram of an apparatus for detecting high frequency components of image data using a conventional average.

Referring to FIG. 1, the apparatus for detecting high frequency components of image data using a conventional average includes a calculation unit 100 for calculating an average value of pixels existing in a window of a surrounding N × N of an input pixel, and an input pixel. If larger than the average, the determination unit 110 determines the pixel as a high frequency pixel. However, the conventional apparatus for detecting high frequency components using the average reflects only the characteristics within the window, and thus there is a problem in that the value is largely changed according to the position of the window. On the other hand, due to such a problem, a pixel determined as a high frequency component may occur in a general image area.

2 is a diagram illustrating the size of the pixel to be irradiated and the size of the surrounding pixel in the conventional method for detecting high frequency components through the size comparison of pixels.

Referring to FIG. 2, it can be seen that the size of the pixel to be irradiated is represented as being much larger than the size of the surrounding pixel (a) or significantly smaller (b). In this case, if the pixel to be investigated is called p (x) and the surrounding pixels are p (x-1) and p (x + 1), the difference between p (x-1) -p (x + 1) is small. When p (x)> (p (x-1) + p (x + 1)) / 2 is satisfied, p (x) is determined as a pixel of high frequency component. However, the conventional high frequency component detection method is advantageous for the detection of the high frequency component of a single pixel, but there is a problem that it is difficult to detect the boundary component appearing in more than one pixel.

SUMMARY OF THE INVENTION The present invention has been made in an effort to provide an apparatus and method for detecting high frequency components of image data capable of expressing a clearer image by detecting high frequency components of image data using distribution characteristics of frequency components of the image data.

1 is a block diagram of a high frequency component detection apparatus of image data using a conventional average;

2 is a diagram illustrating a size of a pixel to be irradiated and a size of a neighboring pixel in a method of detecting a high frequency component by comparing the size of a conventional pixel;

3 is a diagram illustrating a distribution characteristic of image data;

4 is a block diagram showing the configuration of an embodiment of an apparatus for detecting high frequency components of image data according to the present invention;

5 is a flowchart of an embodiment of a method for detecting high frequency components of image data according to the present invention;

6 is a view showing the position and the search order of a window having a size of 50 × 50, and

7A and 7B are diagrams illustrating boundaries and texts that are high frequency components detected by applying the high frequency detection apparatus and method of the original image and the image data according to the present invention, respectively.

In order to achieve the above technical problem, an apparatus for detecting high frequency components of image data according to the present invention includes: a first calculation unit configured to calculate and output an average value of pixels existing in a window of N × N; A second calculation unit for calculating and outputting a predetermined allowable range value according to a frequency of pixel values of the pixels existing in the window; And a detector for detecting a pixel of a high frequency component according to a frequency of pixels outside the allowable range value based on the average value.

Preferably, the tolerance value is a standard deviation of pixel values of the pixels existing in the window. In this case, the detection unit compares the frequency of the pixel having a value larger than the average value plus the allowable range value and the frequency of the pixel having a value smaller than the value obtained by subtracting the allowable range value from the average value to obtain a smaller frequency. A pixel having a high frequency component is determined. In addition, it is preferable to further include a window setting unit for setting the size and / or position of the window.

On the other hand, the high frequency component detection method of the image data according to the present invention comprises the steps of: calculating an average value of the pixels present in the window of N × N; Calculating a predetermined tolerance value according to a frequency of pixel values of pixels present in the window; And detecting a pixel of a high frequency component based on a frequency for pixels deviating from the allowable range value based on the average value.

Preferably, the tolerance value is a standard deviation of pixel values of the pixels existing in the window. In this case, the detecting step compares the frequency of the pixel having a value greater than the average value with the allowable range value and the frequency of the pixel having a value smaller than the value obtained by subtracting the allowable range value from the average value. A pixel having is determined to be a pixel of a high frequency component. The method may further include a window setting step of setting the size and / or position of the window before the average value calculating step.

By using the average value and the statistical characteristics of the peripheral pixels of the pixel to be irradiated, it is possible to accurately detect the high frequency component corresponding to the boundary portion of the image. Also, by keeping the boundary of the image alive, it is possible to prevent an error of mistaken the boundary as noise. Furthermore, the text component can be effectively expressed by preserving the text component detected at the time of scaling the image.

Hereinafter, an apparatus and method for detecting high frequency components of image data according to the present invention will be described in detail with reference to the accompanying drawings.

First, the principle applied to the apparatus and method for detecting high frequency components of image data according to the present invention will be described.

In general, an image can have any value within the range of 0-255. Therefore, the probability that any arbitrary pixel has one of these values is equal to 1/256. Meanwhile, the mean M for any data having a distribution having a random value in the range of 0 to Q and having a probability of 1 / Q is Q × (1 / Q) = 1, and the variance V is to be. The average of these normal RM to the random distribution of expression RM = (maximum value + minimum value) / 2, dispersion RV is (maximum value minimum value) ^2/12.

Analyzing high frequency components using this distribution of data can be said to analyze image data that is not within the range of mean and variance. On the other hand, as shown in Figure 3, the characteristics of the image data can be seen that the frequency of the high frequency component is smaller than the frequency of the general image. Therefore, data having a relatively low frequency that does not fall into the general category can be regarded as a high frequency component.

Figure 4 is a block diagram showing the configuration of an embodiment of an apparatus for detecting high frequency components of image data according to the present invention.

Referring to FIG. 4, the apparatus for detecting high frequency components of image data according to the present invention includes a first operator 400, a second operator 410, a detector 420, and a window setter 430. The first operation unit 400 calculates and outputs an average value of pixels existing in the N × N window. The second operation unit 410 calculates and outputs a standard deviation of pixel values of pixels existing in the N × N window. The standard deviation output from the second operator 410 is a value representing a distribution characteristic of pixels existing in the window, and is a tolerance value used when the detector 420 determines whether the high frequency component is present.

The detector 420 detects a pixel having a high frequency component based on a frequency of pixels out of an allowable range value input from the second operator 410 based on the average value input from the first operator 400. The detector 420 compares the frequency of the pixel having a value larger than the average value plus the allowable range value and the frequency of the pixel having a value smaller than the value obtained by subtracting the allowable range value from the average value. Judging by the pixel of.

The window setting unit 430 sets the size and position of the window represented by N × N. The size of the window set by the window setting unit 430 means a range of peripheral pixels around the pixel to be irradiated. In this case, the plurality of pixels may be selected as the irradiation target pixel. On the other hand, the position of the window is preferably changed so that the detection of the high frequency components can be made for all pixels of the image data.

5 is a flowchart illustrating an embodiment of a method of detecting high frequency components of image data according to the present invention.

Referring to FIG. 5, the window setting unit 430 sets the size and position of the window of N × N by determining the center position of N and the window (S500). The position of the window is determined so that high frequency component detection can be performed on the entire image data or the area to be searched based on the size of the window. For example, as shown in FIG. 6, if the area to be searched is pixels included in a 100 × 100 square in the upper left portion of the image data and the window has a size of 50 × 50, the center of the window is initially set. The position is sequentially changed from the position P to the final position S.

The first operation unit 400 calculates an average value of pixel values of pixels existing in the window (S510), and the second operation unit 410 calculates a standard deviation of pixel values of pixels existing in the window (S520). . The average value and the standard deviation calculated in steps S510 and S520 are for identifying statistical characteristics of pixel values of pixels existing in the window. That is, the process of calculating the allowable range value for removing pixels corresponding to a general image by grasping the frequency of pixel values of pixels existing in a window.

The detector 420 detects a pixel of a high frequency component according to a frequency of pixels deviating from the allowable range value based on the average value (S530). Step S510 has the following substeps. First, the frequency Pmax of pixels satisfying (average value + standard deviation <pixel value) and the frequency Pmin of pixels satisfying (average value-standard deviation> pixel value) are calculated (S532 and S534). Next, the magnitudes of Pmax and Pmin are compared (S536). As a result of the comparison in step S536, when Pmax is greater than Pmin, a pixel satisfying (average value-standard deviation> pixel value) corresponding to Pmin is determined as a high frequency component (S538). On the contrary, when Pmax is smaller than Pmin, pixels satisfying (average value + standard deviation <pixel value) corresponding to Pmax are determined as high frequency components (S540).

The apparatus and method for detecting high frequency components of image data according to the present invention as described above may be employed in an image system for storing and / or transmitting and receiving images. In particular, when employed in the output unit of the imaging system, it is possible to obtain a clearer image by utilizing the boundary portion of the high frequency component when outputting the image. Furthermore, since the text component is a high frequency component, it is possible to effectively express the text included in the image.

7A and 7B are diagrams illustrating boundaries and texts that are high frequency components detected by applying the high frequency detection apparatus and method of the original image and the image data according to the present invention, respectively. Referring to FIGS. 7A and 7B, most text components present in the original image are shown, and the detection degree of the boundary of the image of the original image is high.

According to the apparatus and method for detecting high frequency components of image data according to the present invention, by using the average value and statistical characteristics of the surrounding pixels of the pixel to be irradiated, it is possible to accurately detect the high frequency components corresponding to the boundary of the image, You can get a clearer image by keeping the boundary. In addition, accurate detection of high frequency components can prevent errors that mistake the edges for noise. Furthermore, the text component can be effectively expressed by preserving the text component detected at the time of scaling the image.

Although the present invention has been described in detail through the representative embodiments, those skilled in the art to which the present invention pertains can make various modifications without departing from the scope of the present invention with respect to the embodiments described above. Will understand. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined by the claims below and equivalents thereof.

Claims (8)

A first calculation unit for calculating and outputting an average value of pixels existing in the N × N window; A second calculating unit which calculates and outputs a standard deviation of pixel values of the pixels existing in the window; And And a detector detecting a pixel of a high frequency component according to a frequency of pixels deviating from the standard deviation based on the average value. delete The method of claim 1, The detection unit compares the frequency of a pixel having a value greater than the average value plus the standard deviation and the frequency of a pixel having a value less than the value obtained by subtracting the standard deviation from the average value to obtain a pixel having a smaller frequency. An apparatus for detecting high frequency components of image data, characterized in that it is determined as a pixel of? The method of claim 1, And a window setting unit for setting the size and / or position of the window. Calculating an average value of the pixels present in the N × N window; Calculating a standard deviation of pixel values of pixels present in the window; And And detecting a high frequency component pixel according to a frequency of pixels deviating from the standard deviation based on the average value. delete The method of claim 5, The detecting step compares the frequency of the pixel having a value greater than the average value plus the standard deviation and the frequency of the pixel having a value less than the value obtained by subtracting the standard deviation from the average value. A high frequency component detection method of image data, characterized by judging by a component pixel. The method of claim 5, And a window setting step of setting a size and / or a position of the window before the average value calculating step.