KR101503606B1 - Picture Quality Improvement Apparatus and Method based on Detail Information - Google Patents

Abstract

Disclosed are an apparatus and a method for improving image quality based on the amount of detail information which can improve visual recognition effects in relation to an input image by obtaining detail components from the input image, and improving contrast in proportion to relevant components. The apparatus for improving image quality based on the amount of detail information comprises: an image dividing unit for dividing an input image into various-sized windows for each layer; a detail map constructing unit for constructing a detail map by measuring the amount of detail information of the image divided by the image dividing unit; and an image contrast improving unit for improving image contrast of each layer based on the detail map constructed by the detail map constructing unit; and an image merging unit for merging the image of each layer whose contrast has been improved by the image contrast improving unit with the input image.

Description

[0001] The present invention relates to a picture quality improvement apparatus and a method thereof,

The present invention relates to an image quality improving apparatus and a method thereof, and more particularly, to an image quality improving apparatus and method for improving image quality based on a detail information amount capable of improving a visual cognitive effect on an input image by obtaining a detail component from an input image, An improvement device and a method thereof.

Histogram smoothing is a technique for smoothing the distribution of brightness when the brightness distribution of an image is concentrated. Such histogram smoothing can be roughly divided into global histogram smoothing and local histogram smoothing.

The global histogram smoothing is a method of transforming the distribution of the histogram of the entire image without using the partial region information, and the local histogram smoothing is a method of redistributing the distribution of the histogram using the information of only the corresponding block using the block of a certain size to be.

However, global histogram smoothing is difficult to improve detail contrast of object, and local histogram smoothing can improve detail contrast, but there is a problem that block phenomenon may appear in the result image.

Many studies have been conducted to solve these problems. In this way, there are DSIHE (Dualistic sub-image histogram equalization), POSHE (Partially overlapped sub-block histogram equalization) and MLBOHE (Multiple layer block overlapped histogram equalization).

The above-mentioned three smoothing techniques have a commonality in that the original image is divided into a plurality of blocks or the blocks are used to perform a regional contrast enhancement. N. S. P. Kong proposed [Multiple Layers Block Overlapped Histogram Equalization for Local Content Emphasis, Computers and Electrical Engineering, 37, 2011, pp. 631-643] proposed a MLBOHE technique that divides the original image into three layers and uses the blocks for each layer to equalize the histogram distribution. Then, the original image and contrast enhancement image are merged again to reduce the difference between the original image and the visual image, thereby reducing unnaturalness.

However, the histogram smoothing technique smoothly distributes the concentrated brightness of the image to smooth the visual discrimination of the image. The method described above improves the contrast of the entire image by considering only the brightness distribution within the block without evaluating the original image. There is a problem in that the improvement for the image quality may be insufficient.

Korean Patent Publication No. 2006-126316 Korean Patent Publication No. 2007-048079

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide an image quality improvement device based on a detail information amount capable of improving a visual cognitive effect on an input image by obtaining a detail component from an input image, And to provide such a method.

According to an aspect of the present invention, there is provided an apparatus for improving picture quality based on a detail information amount according to an embodiment of the present invention includes an image divider for dividing an input image into windows of various sizes for each layer; A detail map constructing unit for constructing a detail map by measuring an amount of detail information of the image segmented by the image segmenting unit; An image contrast enhancing unit for enhancing a contrast of a layer-by-layer image based on the detail map configured by the detail map constructing unit; And an image merging unit for merging the input image and the layer-by-layer image, the contrast of which is enhanced by the image contrast enhancing unit.

Wherein the image dividing unit comprises: an image size determining unit for determining a horizontal size M and a vertical size N of the input image, wherein M and N denote an integer as the number of pixels; And a layer classifier for classifying the layer based on the size of the image discriminated by the image size discrimination unit.

The image dividing unit may further include a gray image converting unit for converting gray images into gray images for each window.

The image dividing unit may divide the input image by setting a window size of M / 2 x N / 2, M / 4 x N / 4, and M / 8 x N / 8.

Wherein the detail map constructing unit comprises: a detecting unit that detects a boundary and detail of an image divided by the image dividing unit; A boundary line removal unit for removing an actual boundary line based on the boundary line and the detail detected by the detection unit; And a detail information arithmetic unit for calculating a detail information ratio in a block using a window set in an image in which a boundary line has been removed by the edge line removing unit. According to the present invention, Configure the map.

Wherein the image contrast enhancing unit comprises: a histogram normalizing unit for normalizing a histogram for each block based on a gray image; A probability density calculating unit for calculating a probability density of a histogram for each region based on the histogram normalized by the histogram normalizing unit; A limit setting unit for setting a smoothing limit with respect to a probability density calculated by the probability density calculating unit; And a histogram redistributing unit that distributes the brightness of the block within a range of the threshold set by the limit setting unit.

Wherein the image merging unit comprises: an entropy operation unit for calculating an entropy of an image having improved contrast for each layer; A weight computing unit for computing a weight for each layer based on an entropy calculated by the entropy computing unit; And a layer video merging unit that merges the layer-based video images in proportion to the layer-by-layer video weights calculated by the weight computing unit. The layer video merged by the layer video merging unit and the input video are merged using a predetermined weight .

According to an embodiment of the present invention, there is provided an image quality improvement method based on a detail information amount, comprising: dividing an input image into windows of various sizes for each layer; Measuring a detail information amount of an image segmented by the input image segmenting step to form a detail map; Enhancing the contrast of the layer-by-layer image based on the detail map configured by the detail map constructing step; And merging the input image with the layer-by-layer image having improved contrast by the image contrast enhancement step.

The image quality improvement method includes: determining a horizontal size M and a vertical size N of an input image (where M and N denote an integer as the number of pixels); And classifying the layer based on the size of the image to be discriminated.

The above picture quality improvement method may further include converting each window into a gray image and outputting the converted gray image.

The input image segmenting step may divide the input image by setting a window size of M / 2 x N / 2, M / 4 x N / 4, and M / 8 x N / 8.

The image quality improving method may further include: detecting boundary lines and detail of an image segmented by the input image segmenting step; Removing an actual boundary line based on the boundary line and the detail detected by the detecting step; And calculating a detail information ratio in the block using the window set in the image in which the borderline is removed by the borderline removal step. In this case, the detail map forming step configures a detail map based on the detail information ratio calculated by the detail information ratio calculating step.

Wherein the enhancing of the image contrast comprises: normalizing a histogram for each block based on a gray image; Calculating a probability density of the histogram for each region based on the histogram normalized by the histogram normalization step; Setting a smoothing limit for a probability density calculated by the probability density calculation step; And distributing the brightness of the block within a range of the threshold set by the threshold setting step.

The image quality improvement method includes: calculating an entropy of an image with improved contrast for each layer; Computing a weight for each layer based on entropy calculated by the entropy computing step; And merging the entire layer image in proportion to the layer-by-layer image weight calculated by the weight computing step. In this case, the image merging step merges the layer image and the input image merged by the layer merging step using the set weights.

According to the present invention, a visual cognitive effect on an input image can be improved by obtaining a detail component from an input image and improving contrast in proportion to the component.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 schematically shows a configuration of an image quality improvement apparatus based on a detail information amount according to an embodiment of the present invention; FIG.
Fig. 2 is a view schematically showing the configuration of the image dividing unit shown in Fig. 1; Fig.
3 is a view schematically showing a configuration of a detail map constituting unit shown in Fig.
FIG. 4 is a view schematically showing the configuration of the image contrast enhancing unit shown in FIG. 1; FIG.
5 is a view schematically showing a configuration of a video merge unit shown in Fig.
6 and 7 are flowcharts illustrating an image quality improvement method based on a detail information amount according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, a detailed description of known techniques well known to those skilled in the art may be omitted.

In describing the constituent elements of the present invention, the same reference numerals may be given to constituent elements having the same name, and the same reference numerals may be given thereto even though they are different from each other. However, even in such a case, it does not mean that the corresponding component has different functions according to the embodiment, or does not mean that the different components have the same function. It should be judged based on the description of each component in the example.

In the following description of the embodiments of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

In describing the components of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are intended to distinguish the constituent elements from other constituent elements, and the terms do not limit the nature, order or order of the constituent elements. When a component is described as being "connected", "coupled", or "connected" to another component, the component may be directly connected or connected to the other component, Quot; may be "connected," "coupled," or "connected. &Quot;

FIG. 1 is a view schematically showing a configuration of an image quality improvement apparatus based on a detail information amount according to an embodiment of the present invention.

Referring to FIG. 1, an apparatus 100 for improving picture quality based on a detail information amount according to an embodiment of the present invention includes an image division unit 110, a detail map construction unit 120, an image contrast enhancement unit 130, (140).

The image divider 110 divides the input image into windows of various sizes for each layer. 2, the image divider 110 may include an image size determination unit 112, a layer classification unit 114, and a gray image conversion unit 116. [

The image size determination unit 112 determines the horizontal size M and the vertical size N with respect to the input image. Here, M and N denote integers as the number of pixels. That is, the image size determination unit 112 determines the size of the input image based on the number of pixels in the horizontal direction and the number of pixels in the vertical direction.

The layer classifying unit 114 classifies the layers based on the size of the image discriminated by the image size discriminating unit 112. For example, assuming that a 100.times.100 image is a first layer, the layer classifying unit 114 classifies a 50.times.100 image into a second layer, and a 100.times.50 image into a third layer , And 50 x 50 images can be classified into the fourth layer.

On the other hand, the image divider 110 divides the input image by setting the window size to M / 2 x N / 2, M / 4 x N / 4, and M / 8 x N / 8 for the input image . In the global contrast enhancement step, the images of the divided windows are redistributed to the brightness of the entire image, the details of the boundary line of the object are improved, and the texture expression of the object is improved.

In this case, the gray image converting unit 116 converts the image of each window into a gray image and outputs it.

The detail map constructing unit 120 measures the amount of detail information of the image segmented by the image segmenting unit 110 to construct a detail map. 3, the detail map construction unit 120 may include a detection unit 122, a boundary line removal unit 124, and a detail information calculation unit 126. [

The detection unit 122 detects the boundaries and the detail of the image divided by the image divider 110. At this time, the detector 122 normalizes each pixel level of the gray image to 0 to 1, and if the difference between neighboring pixels of each pixel is greater than a preset threshold value, the detector 122 may generate a binarized image by setting the value of the corresponding pixel as 1. In this case, the binarization image created by Equation (1) includes the detail and boundary line of the object in the image.

[Equation 1]

설정된 임계치를 나타내며, X(i,j)는 그레이 이미지의 I행 j열의 정규화된 밝기 레벨이며, Ex는 구하려는 이진화 영상을 나타낸다.In Equation (1)

X (i, j) represents the normalized brightness level of the I row and j column of the gray image, and Ex represents the binarized image to be obtained.

The boundary line removing unit 124 removes an actual boundary line based on the boundary line and the detail detected by the detecting unit 122. [ That is, the boundary line removing unit 124 estimates the actual boundary line through the edge detection algorithm in the gray image of the previous process, and removes the boundary line from the binarized image.

The detail information calculation unit 126 calculates the detail information ratio in the block using the window set in the image with the borderline removed by the borderline removal unit 124. [

,

의 크기로 설정된다. The detail map constructing unit 120 constructs a detail map based on the detail information ratio calculated by the detail information computing unit 126. That is, the detail map constructing unit 120 may configure the detail map by storing the calculated detail information ratio in the center pixel and performing the process on the entire range of the corresponding layer. The detail map organized by each layer

,

.

The image contrast enhancing unit 130 enhances the contrast of a layer-by-layer image based on a detail map configured by the detail map constructing unit 120. 4, the image contrast enhancement unit 130 may include a histogram normalization unit 132, a probability density calculation unit 134, a limit setting unit 136, and a histogram redistribution unit 138 have.

The histogram normalization unit 132 normalizes the histogram of brightness per block based on the gray image output from the image divider 110. [

The probability density calculation unit 134 calculates the probability density of the histogram for each region based on the histogram normalized by the histogram normalization unit 132. [ The normalized histogram has a value of 1 and can be approximated by Gaussian distribution centering on the mean density region. Equation 2 represents an approximated Gaussian function, where m is the mean density value and? Is the standard deviation.

&Quot; (2) "

The limit setting unit 136 sets a smoothing limit for the probability density calculated by the probability density calculating unit 134. [ That is, the limit setting unit 136 may set the maximum and minimum limits so that the approximated Gaussian function is in inverse proportion to the value stored in the detail map. At this time, the threshold value is set by multiplying by 1/2 according to the characteristic of the detail detection method of detecting the difference from neighboring pixels.

The histogram redistribution unit 138 distributes the brightness of the block within the limit value set by the limit setting unit 136. [ If brightness dispersion of the block is performed on the entire image, an image having improved contrast for each layer is derived.

The image merging unit 140 merges the input image with the layer-by-layer image having improved contrast by the image contrast enhancing unit 130. Since each layer-specific image has a different improved reference, the image merging unit 140 merges the layer image and the input image, and outputs an image in which the contrast between the entire image and the detail object is improved at the same time. 5, the image merging unit 140 may include an entropy computing unit 142, a weight computing unit 144, and a layer image merging unit 146. [

The entropy computing unit 142 computes the entropy of the image with improved contrast for each layer. Equation (3) represents the entropy calculation method, and p (x) is obtained through the process of Equation (4). Here, h (x) represents the histogram of the layer image, and M and N represent the size of the image.

&Quot; (3) "

&Quot; (4) "

The weight computing unit 144 computes a weight for each layer based on the entropy computed by the entropy computing unit 142. [ The amount of information of the corresponding image can be measured through the entropy calculation and the merge weight can be set in proportion to the information amount of the layer image.

The layer video merging unit 146 merges the entire layer video images in proportion to the image weight values of the respective layers calculated by the weight computing unit 144. [ That is, the layer image merging unit 146 merges the entire layer images in proportion to the set image weight per layer, and the image merging unit 140 merges the merged image and the original image using a predetermined weight.

The image output by performing the process like this improves the detail information of the image by analyzing the detail component of the input image and improving the contrast in proportion to the amount of the component. Together with this, the overall contrast of the image can be improved to visually recognize an excessively dark or bright image. This effect can be helpful to users when it is applied to black box or CCTV which is shot at night.

6 and 7 are flowcharts illustrating an image quality improvement method based on the detail information amount according to an embodiment of the present invention. The image quality improvement method based on the detail information amount according to the embodiment of the present invention can be performed by the image quality improvement apparatus 100 described above.

1 to 7, the image divider 110 determines a horizontal size M and a vertical size N with respect to an input image (S102). Here, M and N denote integers as the number of pixels. In addition, the image divider 110 classifies the layers based on the size of the image to be discriminated (S104).

The image divider 110 divides the input image into windows of various sizes for each layer (S106). At this time, the image divider 110 divides the input image by setting the window to a size of M / 2 x N / 2, M / 4 x N / 4, and M / 8 x N / 8 with respect to the input image . In the global contrast enhancement step, the images of the divided windows are redistributed to the brightness of the entire image, the details of the boundary line of the object are improved, and the texture expression of the object is improved.

In this case, the image divider 110 converts the image of each window into a gray image and outputs it (S108).

The detail map constructing unit 120 detects the boundary and detail of the image segmented by the image segmenting unit 110 (S110). At this time, the detail map constructing unit 110 normalizes each pixel level of the gray image to 0 to 1, and if the difference between the neighboring pixels of each pixel is larger than the set threshold value, the detail map constructing unit 110 generates a binarized image .

In addition, the detail map construction unit 120 removes the actual boundary line based on the detected boundary line and detail (S112). That is, the detail map construction unit 120 estimates an actual boundary line through an edge detection algorithm in a gray image of a previous process, and removes a boundary line from the binarized image.

In addition, the detail map construction unit 120 calculates the detail information ratio in the block using the window set in the image with the borderline removed (S114).

,

의 크기로 설정된다.The detail map construction unit 120 constructs a detail map based on the detail information ratio calculated by the detail information calculation unit 126 (S116). That is, the detail map constructing unit 120 may configure the detail map by storing the calculated detail information ratio in the center pixel and performing the process on the entire range of the corresponding layer. The detail map organized by each layer

,

.

The image contrast enhancement unit 130 normalizes the brightness histogram of each block based on the gray image output from the image divider 110 (S118).

In addition, the image contrast enhancement unit 130 calculates the probability density of the histogram for each region based on the normalized histogram (S120). The normalized histogram has a value of 1 and can be approximated by Gaussian distribution centering on the mean density region.

In addition, the image contrast enhancing unit 130 sets a smoothing limit for the calculated probability density (S122). That is, the image contrast enhancement unit 130 may set the maximum and minimum limits so that the approximated Gaussian function is in inverse proportion to the value stored in the detail map. At this time, the threshold value is set by multiplying by 1/2 according to the characteristic of the detail detection method of detecting the difference from neighboring pixels.

In addition, the image contrast enhancing unit 130 distributes the brightness of the corresponding block within the set threshold value (S124). If brightness dispersion of the block is performed on the entire image, an image having improved contrast for each layer is derived.

The image merging unit 140 calculates the entropy of the enhanced image of each layer (S126). In addition, the image merging unit 140 calculates a weight for each layer based on the calculated entropy (S128). The amount of information of the corresponding image can be measured through the entropy calculation and the merge weight can be set in proportion to the information amount of the layer image.

The image merging unit 140 merges the entire layer image in proportion to the calculated image weight per layer (S130). That is, the layer video merging unit 146 merges the entire layer video in proportion to the set video weight per layer.

In addition, the image merging unit 140 merges the merged image and the original image using a predetermined weight (S132).

The present invention is not necessarily limited to these embodiments, as all the constituent elements constituting the embodiment of the present invention are described as being combined or operated in one operation. That is, within the scope of the present invention, all of the components may be selectively coupled to one or more of them. In addition, although all of the components may be implemented as one independent hardware, some or all of the components may be selectively combined to perform a part or all of the functions in one or a plurality of hardware. As shown in FIG. In addition, such a computer program may be stored in a computer-readable medium such as a USB memory, a CD disk, a flash memory, etc., and read and executed by a computer, thereby implementing embodiments of the present invention. As the storage medium of the computer program, a magnetic recording medium, an optical recording medium, a carrier wave medium, or the like may be included.

Furthermore, all terms including technical or scientific terms have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined in the Detailed Description. Commonly used terms, such as predefined terms, should be interpreted to be consistent with the contextual meanings of the related art, and are not to be construed as ideal or overly formal, unless expressly defined to the contrary.

The foregoing description is merely illustrative of the technical idea of the present invention and various changes and modifications may be made by those skilled in the art without departing from the essential characteristics of the present invention. In addition, the embodiments disclosed in the present invention are not intended to limit the scope of the present invention but to limit the scope of the technical idea of the present invention. Accordingly, the scope of protection of the present invention should be construed according to the claims, and all technical ideas within the scope of equivalents should be interpreted as being included in the scope of the present invention.

100: image quality improvement device based on detail information amount,
110: Video division,
120: Detail map constituent part,
130: image contrast enhancement unit,
140: Image merging section.

Claims (14)

An image dividing unit dividing the input image into windows of various sizes for each layer;
A detail map constructing unit for constructing a detail map by measuring an amount of detail information of the image segmented by the image segmenting unit;
An image contrast enhancing unit for enhancing a contrast of a layer-by-layer image based on the detail map configured by the detail map constructing unit; And
And an image merging unit for merging the input image and the layer-by-layer image, the contrast of which is enhanced by the image contrast enhancing unit,
Wherein the image contrast enhancing unit comprises:
A histogram normalization unit for normalizing a histogram for each block based on the gray image;
A probability density calculating unit for calculating a probability density of a histogram for each region based on the histogram normalized by the histogram normalizing unit;
A limit setting unit for setting a smoothing limit with respect to a probability density calculated by the probability density calculating unit; And
A histogram redistribution unit for distributing the brightness of the block within the range of the threshold set by the limit setting unit,
And an amount of detail information based on the amount of detail information.
The method according to claim 1,
Wherein the image dividing unit comprises:
An image size discrimination unit for discriminating a horizontal size M and a vertical size N of the input image (where M and N denote an integer as the number of pixels); And
And a layer classifying unit for classifying the layer based on the size of the image discriminated by the image size discriminating unit,
And an amount of detail information based on the amount of detail information.
3. The method of claim 2,
Wherein the image dividing unit comprises:
A gray image converting unit for converting the gray image into a gray image for each window,
Wherein the detail information amount is based on the detail information amount.
3. The method of claim 2,
Wherein the image dividing unit comprises:
Wherein the window is set to a size of M / 2xN / 2, M / 4xN / 4, and M / 8xN / 8 to divide the input image.
The method according to claim 1,
Wherein the detail map constructing unit comprises:
A detector for detecting a boundary and detail of an image divided by the image divider;
A boundary line removal unit for removing an actual boundary line based on the boundary line and the detail detected by the detection unit; And
A detail information calculation unit for calculating a detail information ratio in a block using a window set in an image in which a boundary line has been removed by the boundary line removal unit;
/ RTI >
And a detail map is formed based on the detail information ratio calculated by the detail information calculation unit.
delete The method according to claim 1,
Wherein the image merging unit comprises:
An entropy computing unit for computing an entropy of an image having an improved contrast for each layer;
A weight computing unit for computing a weight for each layer based on an entropy calculated by the entropy computing unit; And
And a layer image merging unit for merging the entire layer images in proportion to the image weight values of the layers calculated by the weight calculating unit,
/ RTI >
And merging the layer video and the input video merged by the layer video merging unit using a predetermined weight value.
A picture quality improvement method based on the amount of detail information performed by the picture quality improvement apparatus,
Dividing an input image into windows of various sizes for each layer;
Measuring a detail information amount of an image segmented by the input image segmenting step to form a detail map;
Enhancing the contrast of the layer-by-layer image based on the detail map configured by the detail map constructing step; And
Merging an input image and a layer-by-layer image having improved contrast by the image contrast enhancing step,
The image contrast enhancing step may include:
Normalizing the per-block histogram based on the gray image;
Calculating a probability density of the histogram for each region based on the histogram normalized by the histogram normalization step;
Setting a smoothing limit for a probability density calculated by the probability density calculation step; And
Distributing the brightness of the block within a range of the threshold value set by the threshold setting step
And generating a detail information amount based on the amount of detail information.
9. The method of claim 8,
Determining a horizontal size M and a vertical size N of the input image, wherein M and N denote an integer as the number of pixels; And
Classifying the layer based on the size of the image to be discriminated
And generating a detail information amount based on the detail information amount.
9. The method of claim 8,
Converting each window into a gray image and outputting
And generating a detail information amount based on the detail information amount.
9. The method of claim 8,
Wherein the input image segmenting step comprises:
Wherein the input image is divided by setting a window size of M / 2 x N / 2, M / 4 x N / 4, and M / 8 x N / 8.
9. The method of claim 8,
Detecting boundary lines and details of an image segmented by the input image segmenting step;
Removing an actual boundary line based on the boundary line and the detail detected by the detecting step; And
Calculating a detail information ratio in the block using the window set in the image with the border removed by the border removing step
Further comprising:
Wherein the detail map forming step configures the detail map based on the detail information ratio calculated by the detail information ratio calculating step.
delete 9. The method of claim 8,
Calculating an entropy of an image in which the contrast of each layer is improved;
Computing a weight for each layer based on entropy calculated by the entropy computing step; And
Merging the entire layer image in proportion to the image weight per layer calculated by the weight computing step
Further comprising:
Wherein the merging step merges the layer image and the input image merged by the merging step of the whole layer image using a predetermined weight value.

Images