KR101854611B1 - Image Processing Method Based on Weight Assignment using Entropy - Google Patents

Abstract

The present invention relates to an image processing method using weight allocation using entropy, and more particularly, to an image processing method for allocating a weight using information entropy while converting an image with a progressive frame rate by reconfiguring an interlace signal in a vertical direction by an interpolation method in consideration of an edge direction. Accordingly, the presetn invention can improve image quality.

Description

[0001] IMAGE PROCESSING METHOD BASED ON WEIGHT ALLOCATION BY USING ENTROPY [0002]

The present invention relates to an image processing method, and more particularly, to a new image processing method for converting an image at an interlaced to a progressive frame rate.

Video signal processing uses space and time information to recover missing information. This information may be used separately or in combination. Interlace signals have been used for a long time. Many issues have been researched, but there are still issues to be addressed.

As one of the simplest methods, there is a conventional method Bob for recovering missing lines between two neighboring pixels present in an interlaced image before displaying an image. While temporal information based methods exhibit better performance, the complexity is usually higher than spatial methods because of the need for motion models and tracking methods. However, since only spatial domain methods use a given field, these methods are widely used in many cases for simplification.

One of the widely used methods is the well known edge-based line averaging method which utilizes correlation in three directions between upper and lower lines of pixels. Recently, an efficient edge direction line averaging method has been proposed.

However, there is a need for a more effective way of converting from interlaced to progressive frame rates.

SUMMARY OF THE INVENTION Accordingly, the present invention has been made keeping in mind the above problems occurring in the prior art, and it is an object of the present invention to provide an image processing apparatus and method, and an image processing method for assigning weights using information entropy.

According to another aspect of the present invention, there is provided an image processing method for an image processing apparatus, the method comprising the steps of: Calculating a pixel value of the first method; Using the neighboring pixel values in the upper and lower lines and the neighboring pixel values in the diagonal direction corresponding to the upper and lower lines and the neighboring pixel values in the upper and lower lines and the neighboring pixel values in the other diagonal direction corresponding to the neighboring pixel values in the upper and lower lines, Calculating a pixel value of a second method for the second method; Generating an entropy map from the interlace signal by calculating an entropy pixel value based on a probability of becoming an actual pixel value for each of the target pixels; And assigning a weight to each pixel value of the first method and the second method using the entropy map to determine a final pixel value.

A value corresponding to each pixel of the entropy map has a value between 0 and 1.

The step of determining the final pixel value may further include assigning a larger weight value to the pixel value calculated according to the first method when the value corresponding to each pixel of the entropy map is close to 1 and corresponding to each pixel of the entropy map And giving a larger weight to the pixel value calculated according to the second method if the value is close to zero.

Wherein the step of determining the final pixel value comprises the steps of determining a pixel value calculated according to the first method as the final pixel value if the value corresponding to each pixel of the entropy map is greater than a threshold value, And determining the pixel value calculated according to the second method as the final pixel value if the corresponding value is below the threshold value.

In the step of generating the entropy map, the entropy pixel value H (A)

, Where n is the number of random variables, b is the base of the log, and P (a _i ) is the probability for a random variable (a _i ) associated with pixel values of neighboring upper and lower line pixels at each target pixel .

A recording medium implemented with a computer readable code according to another aspect of the present invention includes a function of calculating a pixel value of a first method for each target pixel using an average value of pixel values of upper and lower lines from an interlace signal ; Using the neighboring pixel values in the upper and lower lines and the neighboring pixel values in the diagonal direction corresponding to the upper and lower lines and the neighboring pixel values in the upper and lower lines and the neighboring pixel values in the other diagonal direction corresponding to the neighboring pixel values in the upper and lower lines, Calculating a pixel value of a second method for the pixel; A function of generating an entropy map from the interlace signal by calculating an entropy pixel value based on a probability of becoming an actual pixel value for each of the target pixels; And assigning a weight to each pixel value of the first method and the second method using the entropy map to determine a final pixel value.

According to the image processing method according to the present invention, it is possible to reconstruct an interlaced signal in a vertical direction to convert the image into a progressive image by using information entropy in a conventional line recovery method (Bob) and a conventional efficient edge direction line average (EELA) ) Is used to assign a weight value, there is an effect of improving image quality such as sharpness improvement compared with the conventional method.

1 is a diagram for explaining a conventional edge direction line averaging method.
FIG. 2 is a diagram illustrating a conventional efficient edge-directed line average (EELA) method. Referring to FIG.
3 is a diagram for explaining the relationship between the original image and the entropy map of the present invention.
4 is a flowchart illustrating an image processing method according to an embodiment of the present invention.
FIG. 5 is a diagram showing an image processing result (c, d) by the conventional methods, an entropy map (b) according to the present invention and an image processing result (e) (F, g, h) between the image and the original image.
Fig. 6 is an example of other images for comparing the image processing results (b, c) and the image processing result (d) of the present invention with the original input image (a) by the conventional methods.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same components are denoted by the same reference symbols as possible. In addition, detailed descriptions of known functions and / or configurations are omitted. The following description will focus on the parts necessary for understanding the operation according to various embodiments, and a description of elements that may obscure the gist of the description will be omitted. Also, some of the elements of the drawings may be exaggerated, omitted, or schematically illustrated. The size of each component does not entirely reflect the actual size, and therefore the contents described herein are not limited by the relative sizes or spacings of the components drawn in the respective drawings.

1 is a diagram for explaining a conventional edge direction line averaging method.

Where P and Q are dominant edge direction selectors.

1, a predetermined edge detector calculates pixel values k (x, y) in the diagonal direction in the upper and lower lines with respect to the target pixel (x, y) to be interpolated and estimated as shown in Equations (1) (X-1, y-1), k (x + 1, y + 1) Q can be calculated.

[Equation 1]

&Quot; (2) "

FIG. 2 is a diagram illustrating a conventional efficient edge-directed line average (EELA) method. Referring to FIG.

Referring to FIG. 2, in the conventional efficient edge direction line averaging method, the predominant edge direction selectors P and Q are used in two diagonal directions using two neighboring pixel values of two lines as shown in [Equation 3] and [Equation 4] To reflect the four pixel values.

&Quot; (3) "

&Quot; (4) "

The entropy referred to in the present invention is an evaluation measure for order and disorder in information theory. In the present invention, entropy is used as a weighting metric for information. For example, the entropy H (A) for the event A according to Shannon's entropy concept can be expressed by using the probability P (a _i ) for a certain random variable a _i as shown in [Equation 5] Can be represented by the sum of states. Where n is the number of random variables, b is the basis of the log, and b can be 2, 10, and so on.

&Quot; (5) "

3 is a diagram for explaining the relationship between the original image and the entropy map of the present invention.

In the present invention, when H (A) is viewed as an entropy pixel value at a target pixel (x, y) (corresponding to event A), a map for H (A) (see Fig. 3 (b)). The entropy pixel value at each pixel has a value between the lowest gray (0, e.g., black) and the maximum gray (1, e.g., white). P (a _i) is a probability for each target pixel (x, y) pixel values of the random variable (a _i) associated with the pixels of the upper and lower lines adjacent in the probability P (a _i) is each of the pixels (x, (x, y + 1), k (x, y-1)) of the pixels of the upper and lower lines, and the like. The probability P (a _i ) is calculated by multiplying the pixel values of the pixels of the upper and lower lines adjacent to each other in the target pixel (x, y), for example, k (x, y + (x-1, y + 1), k (x + 1, y-1) k (x-1, y-1), and so on.

4 is a flowchart illustrating an image processing method according to an embodiment of the present invention.

Hereinafter, the edge direction is evaluated and reflected by applying the conventional line recovery method (Bob) and the conventional efficient edge direction line average (EELA) method in converting the interlaced signal in the vertical direction into the progressive image and reflecting the information entropy information a new image processing method according to an embodiment of the present invention for enhancing image quality such as sharpness improvement compared to the conventional method will be described by allocating weights using entropy.

The new image processing method according to an embodiment of the present invention can be implemented in a predetermined image processing apparatus, and the image processing apparatus can be implemented by hardware such as a semiconductor processor, software such as an application program, or a combination thereof . The image processing device may be a computing system and may include at least one processor, a memory, a user interface input device, a user interface output device, a storage, and a network interface. A processor may be a central processing unit (CPU) or a semiconductor device that performs processing on instructions stored in memory or storage.

In the image processing apparatus of the present invention, when a color image (data) (for example, R (Red) G (Green) (Y) and a YCbCr color image (data) composed of color difference data (Cb, Cr) (S10).

The image processing apparatus decimates the original luminance data Y (S10). The image processing apparatus decimates the original luminance data Y in the vertical direction of the image to generate an interlace signal (or data). For example, it is possible to repeat the process of selecting the odd lines of the first field in the original luminance data Y and selecting the even lines in the next field. An interlace signal can be generated for the color difference data (Cb, Cr) in the same manner.

Next, the image processing apparatus first reconstructs the interlaced signal in the vertical direction by the conventional line restoration method (Bob) S20 and the efficient edge direction line average (EELA) method (S30, S31), respectively, and converts it into a progressive signal .

For example, in the conventional line recovery method (Bob) S20, the average value of the pixel values k (x, y + 1) and k (x, y- Can be calculated as a pixel value. For example, in FIG. 5, the image processing result according to the conventional line recovery method (Bob) is as shown in FIG. 5 (c).

In an efficient edge direction line average (EELA) method (S30, S31), neighboring pixel values of upper and lower lines (e.g., k (x-1, y) 1), k (k + 1, y + 1), etc.) and corresponding diagonal neighboring pixel values (e.g., k The parameter P is calculated and the neighboring pixel values of the upper and lower neighboring pixels in the other diagonal direction (e.g., k (x-1, y + 1), k (X, y) by assigning appropriate weights to P and Q by calculating the parameter Q by using the above equation (e.g., k (x-1, y), k The pixel value can be calculated. For example, in FIG. 5, the image processing result according to the conventional efficient edge direction line average (EELA) method is as shown in FIG. 5 (d).

In the present invention, the image processing apparatus generates an entropy map from the interlace signal using Equation (5) above (S40). The entropy pixel values H (A) at the target pixel (x, y) (corresponding to the event A) are calculated from the corresponding image having the interlace signal so that the entropy pixel value at each pixel is the lowest gray (0, An entropy map (see (e) in FIG. 5) having a value between maximum gray (1, e.g., white) can be generated. The probability P (a _i) from the equation (5)] is the target pixel (x, y) of the pixel values of the pixels, for example, k (x, y + 1) neighboring a, k (x, y-1 ) (x-1, y + 1), k (x + 1, y), k ), k (x-1, y-1), ..., and so forth.

The image processing apparatus then uses the entropy map to assign weights to the pixel values of the target pixel x, y according to the results of the conventional method 1 (Bob) and method 2 (EELA) And outputs an image (S50).

For example, if the entropy pixel value of the entropy map for the target pixel (x, y) is close to 1, give a larger weight to the pixel value calculated according to method 1 (Bob), and if the entropy pixel value is close to zero A larger weight can be given to the pixel value calculated according to Method 2 (EELA).

If the entropy pixel value is larger than a predetermined threshold value (for example, 0.1 to 0.7) as in the following algorithm, the pixel value calculated according to the method 1 (Bob) is selected and the target pixel x, y ), And if the entropy pixel value is less than or equal to a predetermined threshold value (e.g., 0.1 to 0.7), the pixel value calculated according to the method 2 (EELA) is selected and the target pixel x, y ≪ / RTI > as a final pixel value.

 [algorithm]

FIG. 5 is a diagram showing an image processing result (c, d) by the conventional methods, an entropy map (b) according to the present invention and an image processing result (e) (F, g, h) between the image and the original image. As shown in FIG. 5E, the image processing result using the entropy map (b) according to the present invention is obtained by the image processing result (c) by the conventional method 1 (Bob) and the image processing result by the conventional method 2 (EELA) It is confirmed that the image processing result is clearer than the result (d) and the edge part is well maintained. This is because the difference between the original input image (a) and the image processing result (e) of the present invention compared to the difference f / g between the original input image (a) and the conventional method 1 (Bob) / method 2 h can be seen from the fact that the edge portion is further emphasized.

As a result of evaluating the image quality by calculating the peak signal-to-noise ratio (PSNR) for each of R (Red), G (Green) and B It is confirmed that the PSNR result is more appropriate than the other conventional methods.

6, the conventional method 1 (Bob) (b) for the input image (a) shows staircase artifacts in the line than the method 2 (EELA) (c) And the image blur resulting from the image processing result (d) of the present invention can obtain clear images with improved noise compared to other methods.

As described above, in the image processing method in the image processing apparatus according to the present invention, the conventional line recovery method (Bob) and the conventional efficient edge direction line average (RB) in the vertical direction reconstruction by converting the interlace signal into the progressive image EELA) method by assigning weights by using information entropy, there is an effect of improving image quality, such as improving the sharpness, compared with the conventional method.

The functions for performing the image processing method using weighting using entropy in the image processing apparatus according to an embodiment of the present invention include a computer readable code on a recording medium readable by an apparatus such as a computer, A combination of such a recording medium and a device such as a computer can be implemented to input, output, and display data or information necessary for performing a function. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, hard disk, removable storage device and the like.

As described above, the present invention has been described with reference to particular embodiments, such as specific elements, and specific embodiments and drawings. However, it should be understood that the present invention is not limited to the above- Those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the essential characteristics of the invention. Therefore, the spirit of the present invention should not be construed as being limited to the embodiments described, and all technical ideas which are equivalent to or equivalent to the claims of the present invention are included in the scope of the present invention .

Claims (10)

A method of processing an image in an image processing apparatus,
Calculating a pixel value of a first method for each target pixel using an average value of pixel values of the upper and lower lines from an interlaced signal;
Using the neighboring pixel values in the upper and lower lines and the neighboring pixel values in the diagonal direction corresponding to the upper and lower lines and the neighboring pixel values in the upper and lower lines and the neighboring pixel values in the other diagonal direction corresponding to the neighboring pixel values in the upper and lower lines, Calculating a pixel value of a second method for the second method;
Generating an entropy map from the interlace signal by calculating an entropy pixel value based on a probability of becoming an actual pixel value for each of the target pixels; And
And assigning a weight to each pixel value of the first method and the second method using the entropy map to determine a final pixel value,
A value corresponding to each pixel of the entropy map has a value between 0 and 1,
In determining the final pixel value,
When a value corresponding to each pixel of the entropy map is close to 1, a larger weight is given to a pixel value calculated according to the first method, and when a value corresponding to each pixel of the entropy map is close to 0, A larger weight is given to the pixel value calculated according to the method,
If the value corresponding to each pixel of the entropy map is greater than a predetermined threshold value between 0.1 and 0.7, the pixel value calculated according to the first method is determined as the final pixel value, and a value corresponding to each pixel of the entropy map Is less than or equal to the threshold value, the pixel value calculated according to the second method is determined as the final pixel value. delete delete delete The method according to claim 1,
In the step of generating the entropy map,
The entropy pixel value H (A)

, Where n is the number of random variables, b is the base of the log, P (a _i ) is the probability for a random variable (a _i ) associated with pixel values of neighboring upper and lower line pixels at each target pixel The image processing method comprising the steps of: A function of calculating a pixel value of a first method for each target pixel using an average value of pixel values of upper and lower lines from an interlaced signal;
Using the neighboring pixel values in the upper and lower lines and the neighboring pixel values in the diagonal direction corresponding to the upper and lower lines and the neighboring pixel values in the upper and lower lines and the neighboring pixel values in the other diagonal direction corresponding to the neighboring pixel values in the upper and lower lines, Calculating a pixel value of a second method for the pixel;
A function of generating an entropy map from the interlace signal by calculating an entropy pixel value based on a probability of becoming an actual pixel value for each of the target pixels; And
Performing a function of assigning weights to pixel values of the first method and the second method using the entropy map to determine a final pixel value,
A value corresponding to each pixel of the entropy map has a value between 0 and 1,
In the function of determining the final pixel value,
When a value corresponding to each pixel of the entropy map is close to 1, a larger weight is given to a pixel value calculated according to the first method, and when a value corresponding to each pixel of the entropy map is close to 0, A larger weight is given to the pixel value calculated according to the method,
If the value corresponding to each pixel of the entropy map is greater than a predetermined threshold value between 0.1 and 0.7, the pixel value calculated according to the first method is determined as the final pixel value, and a value corresponding to each pixel of the entropy map Is less than or equal to the threshold value, the pixel value calculated according to the second method is determined as the final pixel value. delete delete delete The method according to claim 6,
In the function of generating the entropy map,
The entropy pixel value H (A)

, Where n is the number of random variables, b is the base of the log, P (a _i ) is the probability for a random variable (a _i ) associated with pixel values of neighboring upper and lower line pixels at each target pixel .

Images