WO2017007047A1

WO2017007047A1 - Spatial depth non-uniformity compensation method and device using jittered comparison

Info

Publication number: WO2017007047A1
Application number: PCT/KR2015/007044
Authority: WO
Inventors: 이영민; 윤우진; 박현상; 경종민
Original assignee: 재단법인 다차원 스마트 아이티 융합시스템 연구단
Priority date: 2015-07-08
Filing date: 2015-07-08
Publication date: 2017-01-12

Abstract

A spatial depth non-uniformity compensation method using jittered comparison comprises the steps of: acquiring a first image and a second image; detecting edges from each of the first image and the second image; performing the jittered comparison for pixels of the edges on the basis of an offset jitter map in which an offset jitter vector corresponding to each of sub-regions divided from the first image and the second image is stored; and determining a depth for the pixels of the edges for which the jittered comparison has been performed, by using a compensation model in which a depth corresponding to a blur size corresponding to each of the sub-regions is stored.

Description

Method and apparatus for compensating spatial depth non-uniformity using irregular comparison

The embodiments below relate to a method and apparatus for spatial depth non-uniformity compensation using jittered comparison, and more particularly blur size at the pixel of the edge detected in two different images. In the process of determining the depth corresponding to, the technique of performing spatial depth non-uniformity compensation using an irregular comparison.

The blurring of the pixels of the edge detected in the image is expressed as a function of the size of the aperture included in the camera system, the distance between the subject and the camera system, and the focal length of the lens included in the camera system. The technique of determining the depth of the depth is determined based on the relationship of the blur size to the pixel of the edge detected in each of the two different images.

However, the techniques for determining the depth of an existing image are each different due to defect imperfections of an aperture lens system such as chromatic aberration, spherical aberration or aperture off-axis in the process of generating two different images. There is a disadvantage in that the edges detected in each of the other two images are inconsistent with each other.

In addition, the conventional technique for determining the depth of an image has a problem that a depth determination error occurs due to a spatial nonuniformity in which the blur size between the center pixel and the edge pixel of the image is changed.

Accordingly, in the present specification, in the process of determining the depth of an image, a technique of performing random comparison and spatial depth nonuniformity compensation to match edges detected in each of two different images and solving spatial nonuniformity is proposed.

The following embodiments provide a method and apparatus for matching edges detected in each of two different images and resolving spatial nonuniformity in determining a depth of an image.

Specifically, the following embodiments provide a method and apparatus for performing spatial depth non-uniformity compensation using irregular comparisons to match edges detected in each of two different images and to resolve spatial non-uniformity.

According to one embodiment, a spatial depth non-uniformity compensation method using jittered comparison includes obtaining a first image and a second image; Detecting an edge in each of the first image and the second image; Performing an irregular comparison on a pixel of the edge based on an offset jitter map in which an offset irregular vector corresponding to each of the sub-regions in which the first image and the second image are divided is stored; And determining a depth of a pixel of the edge on which the irregular comparison is performed by using a compensation model in which a depth according to a blur size corresponding to each of the subregions is stored.

Performing an irregular comparison with the pixels of the edges may include setting a first patch in a preset range around the pixels of the edges detected in the first image; Selecting an offset irregular vector corresponding to a subregion including a pixel of the edge detected in the first image from the offset irregular map; And obtaining a second patch having the highest similarity to the first patch within a preset jitter range around a pixel shifted by the selected offset irregular vector from the pixel of the edge detected in the second image. It may include a step.

The acquiring of the second patch may include acquiring the second patch centered on a pixel shifted by the delta irregular vector from the selected offset irregular vector.

Determining a depth for the pixel of the edge on which the irregular comparison is performed comprises extracting a depth corresponding to the blur size in the first patch and the second patch using the compensation model; And determining the extracted depth as a depth of a pixel of the edge on which the irregular comparison is performed.

The spatial depth non-uniformity compensation method may further include forming the offset irregular map.

The forming of the offset irregular map may include setting a first patch in a preset range around a pixel of the edge detected in the first image; Acquiring a third patch having the highest similarity to the first patch within a preset search area around the pixels of the edge detected in the second image; Calculating a distance vector between the first patch and the third patch; And statistically setting an offset irregular vector corresponding to each of the subregions based on the distance vector calculated for the pixel of the edge included in each of the subregions.

The spatial depth nonuniformity compensation method may further include constructing the compensation model.

The constructing of the compensation model may include setting a depth according to a blur size corresponding to each of the subregions in consideration of a center pixel and an edge pixel in each of the subregions.

According to one embodiment, a spatial depth non-uniformity compensation apparatus using jittered comparison includes an image acquisition unit for acquiring a first image and a second image; An edge detector detecting an edge in each of the first image and the second image; Performing an irregular comparison on a pixel of the edge based on an offset jitter map storing an offset irregular vector corresponding to each of the subregions in which the first image and the second image are divided; part; And a depth determiner configured to determine a depth of a pixel of the edge on which the irregular comparison is performed, using a compensation model in which a depth according to a blur size corresponding to each of the subregions is stored.

The irregular comparison performing unit sets a first patch having a preset range around a pixel of the edge detected in the first image, and includes a pixel of the edge detected in the first image from the offset irregular map. Selecting an offset irregular vector corresponding to an area, wherein the first patch is located within a preset jitter range about a pixel shifted from the pixel of the edge detected in the second image by the selected offset irregular vector. A second patch having the highest similarity can be obtained.

The depth determiner extracts a depth corresponding to the blur size in the first patch and the second patch by using the compensation model, and converts the extracted depth into a depth of a pixel of the edge on which the irregular comparison is performed. You can decide.

The spatial depth non-uniformity compensation device may further include an offset irregular map forming unit forming the offset irregular map.

The offset irregular map forming unit sets a first patch in a preset range around a pixel of the edge detected in the first image, and within a preset search area around the pixel of the edge detected in the second image. Obtains a third patch having the highest similarity to the first patch, calculates a distance vector between the first patch and the third patch, and calculates a pixel of an edge included in each of the subregions. An offset irregular vector corresponding to each of the subregions may be statistically set based on the calculated distance vector.

The spatial depth non-uniformity compensation device may further include a compensation model building unit for building the compensation model.

The compensation model builder may set a depth according to a blur size corresponding to each of the subregions in consideration of a center pixel and an edge pixel in each of the subregions.

The following embodiments may provide a method and apparatus for matching edges detected in each of two different images and resolving spatial nonuniformity in determining a depth of an image.

Specifically, the following embodiments may provide a method and apparatus for performing spatial depth non-uniformity compensation using irregular comparison to match edges detected in each of two different images, and to solve spatial non-uniformity. .

1 is a diagram illustrating subregions in which an image is segmented to perform spatial depth non-uniformity compensation using irregular comparison, according to an exemplary embodiment.

FIG. 2 is a diagram illustrating a process of forming an offset irregular map in order to perform spatial depth non-uniformity compensation using irregular comparison, according to an exemplary embodiment.

3 illustrates an offset irregular map formed according to an exemplary embodiment.

4 illustrates a compensation model constructed according to an embodiment.

5 is a diagram for describing a process of performing spatial depth non-uniformity compensation using irregular comparison, according to an exemplary embodiment.

6 is a flow chart illustrating a spatial depth non-uniformity compensation method using an irregular comparison according to an embodiment.

7 is a block diagram illustrating a spatial depth non-uniformity compensation device using an irregular comparison according to an embodiment.

Hereinafter, exemplary embodiments will be described in detail with reference to the accompanying drawings. However, the present invention is not limited or limited by the embodiments. Also, like reference numerals in the drawings denote like elements.

Also, the terminology used herein is a term used to properly express a preferred embodiment of the present invention, which may vary depending on a user, an operator's intention, or customs in the field to which the present invention belongs. Therefore, the definitions of the terms should be made based on the contents throughout the specification.

Referring to FIG. 1, the spatial depth non-uniformity compensation apparatus according to an embodiment acquires a first image and a second image, and then performs a first image and a second image to perform spatial depth non-uniformity compensation using an irregular comparison. The entire region of may be divided into M × N sub-regions 110. At this time, M≥1 & N≥1, and the unit may be the number of pixels.

The spatial depth non-uniformity compensation device may apply an irregular comparison as shown in Equation 1 to the divided subregions 110.

[Equation 1]

J _total (x, y) = J _offset ^{i (x, y)} + J _delta ^{i (x, y)} (x, y)

In Equation 1, x represents the x-axis of the pixel, y represents the y-axis of the pixel, J _total (x, y) represents the irregular vector of the pixel of the edge having the x, y position, J _offset ^{i (x, y)} denotes an offset irregular vector corresponding to the subregion i (x, y) where the pixels of the edge are included, and J _delta ^{i (x, y)} (x, y) denotes the subregion i ( delta delta irregular vector from the offset irregular vector corresponding to x, y).

Therefore, the spatial depth non-uniformity compensation device sets and stores the offset irregular vector J _offset ^{i (x, y)} corresponding to each of the subregions 110 before performing the random comparison, thereby offset offset jitter map ) Is formed in advance and an offset irregular vector corresponding to the subregion i (x, y) including the pixel of the edge of the subregions 110 in the process of performing the irregular comparison is J _delta ^{i (x, y)} ( By calculating x, y), an irregular comparison can be performed on the pixels at the edges. Detailed description thereof will be described below.

Referring to FIG. 2, the spatial depth non-uniformity compensation device according to an embodiment may preform an offset irregular map before performing spatial depth non-uniformity compensation using an irregular comparison.

Specifically, the spatial depth non-uniformity compensation device detects an edge in the first image 210 and the second image 220, and then sets a predetermined range around the pixels 212 of the edge detected in the first image 210. The patch 211 of the first image 210 may be set.

Then, the spatial depth non-uniformity compensator preliminarily centers on the pixel 222 of the edge detected in the second image 220 (the pixel at the same position as the pixel 212 of the edge detected in the first image 210). The patch 221 of the second image 220 having the highest similarity with the patch 211 of the first image 210 may be acquired within the set jitter range 230.

Thereafter, the spatial depth non-uniformity compensation device may calculate the distance vector jitter 240 between the patch 211 of the first image 210 and the patch 221 of the second image 220.

In this case, the spatial depth non-uniformity compensating apparatus may perform subregions based on the distance vector 240 calculated for the pixel of the edge included in each of the subregions in which the first image 210 and the second image 220 are divided. The offset irregular vectors corresponding to each can be statistically set.

For example, the spatial depth non-uniformity compensation device calculates the distance vector 240 of each of the pixels of the edge in each of the subregions in which the first image 210 and the second image 220 are divided, and then the subregions. An offset irregular vector corresponding to each of the subregions may be set as an average value of the distance vectors 240 of the pixels of the edges obtained in each of the subregions.

Here, the spatial depth non-uniformity compensation device reduces the time and complexity of the process of forming the offset irregular map, instead of calculating the distance vector 240 of all the pixels of the edge in each of the subregions to obtain an average value. The average value may be obtained by calculating a distance vector 240 of some pixels of the edge pixels in each of the subregions.

When the offset irregularity vector corresponding to each of the subregions is set as described above, the spatial depth non-uniformity compensation device determines the depth by storing and maintaining an offset irregular map in which the offset irregularity vector corresponding to each of the subregions is stored in the memory. Irregular comparisons can be performed at. Detailed description thereof will be described with reference to FIG. 3.

In addition, the spatial depth non-uniformity compensation device may build a compensation model in advance before performing the spatial depth non-uniformity compensation in the process of determining the depth.

For example, the spatial depth non-uniformity compensation device sets the depth according to the blur size corresponding to each of the sub-areas in consideration of the center pixel and the edge pixel in each of the sub-areas, thereby reducing the blur size corresponding to each of the sub-areas. The compensation model stored according to the depth can be stored in the memory and maintained. Detailed description thereof will be described with reference to FIG. 4.

Referring to FIG. 3, the spatial depth non-uniformity compensating device according to an exemplary embodiment of the present invention may further include each of the subregions based on a distance vector calculated for a pixel of an edge included in each of the subregions in which the first and second images are divided. By statistically setting the offset irregularity vector corresponding to, the offset irregularity map 310 may be formed. Here, the offset irregular vector corresponding to each of the subregions may be composed of an offset of the x-axis coordinate of the pixel and an offset of the y-axis coordinate of the pixel.

In this case, the offset irregular map 310 may be formed in advance in the process of producing the spatial depth non-uniformity compensation device (in particular, during production inspection).

4 illustrates a compensation model constructed according to an embodiment.

Referring to FIG. 4, the spatial depth non-uniformity compensating apparatus according to an exemplary embodiment may determine the depth according to the blur size corresponding to each of the subregions in consideration of the center pixel 411 and the edge pixel 412 in each of the subregions. The compensation model can be constructed by setting the transformation curve 410 (in the transformation curve 410, the x-axis represents depth and the y-axis represents the blur size).

For example, the spatial depth non-uniformity compensation device may set the depth according to the blur size corresponding to (1 × 1), which is the first subregion, to the first transformation curve 420, and (1 × 2, which is the second subregion. Depth according to the blur size corresponding to) may be set to the second conversion curve equation 430.

At this time, the compensation model, like the offset irregular map, may be built in advance in the process of producing the spatial depth non-uniformity compensation device (especially during production inspection).

Referring to FIG. 5, a spatial depth non-uniformity compensating apparatus detects edges in a first image 510 and a second image 520, and then detects edges of pixels of an edge based on a preformed offset irregular map. Perform random comparisons.

In detail, the apparatus for compensating for spatial depth non-uniformity may set the patch 511 of the first image 510 in a preset range around the pixel 512 of the edge detected in the first image 510.

Subsequently, the spatial depth non-uniformity compensator may select an offset irregular vector 530 corresponding to a subregion including the pixel 512 of the edge detected in the first image 510 from the previously formed offset irregular map.

Then, the spatial depth non-uniformity compensation device selects an offset irregularity selected from the pixel 522 of the edge detected in the second image 520 (the pixel at the same position as the pixel 512 of the edge detected in the first image 510). The patch 521 of the second image 520 that is most similar to the patch 511 of the first image 510 within the preset search area 540 around the pixel 531 moved by the vector 530. Can be obtained.

In this case, since the delta irregular vector 550 represents the vector distance from the selected offset irregular vector 530 to the pixel 523 which is the center of the patch 521 of the second image 520, the second image 520 The process of acquiring a patch 521 is calculated by calculating a delta irregular vector 550 and obtaining a patch centered on the pixel 523 moved by the delta irregular vector 550 calculated from the selected offset irregular vector 530. It may be a process of obtaining.

After the irregular comparison is performed on the pixels of the edge through this process, the spatial depth non-uniformity compensation device determines the depth of the pixels of the edge on which the irregular comparison is performed, using a previously constructed compensation model.

Specifically, the spatial depth non-uniformity compensation device extracts a depth corresponding to the blur size in the patch 511 of the first image 510 and the patch 521 of the second image 520 using a pre-built compensation model. Then, the extracted depth may be determined as the depth for the pixel of the edge where the irregular comparison is performed.

Therefore, the spatial depth non-uniformity compensation device not only performs the irregular comparison and spatial depth non-uniformity compensation described above with respect to the other pixels of the edge detected in the first image 510 and the second image 520, but also the first image. By performing the 510 and the second image 520 on each of the pixels included in the divided subregions, a depth of each of the subregions may be determined.

In addition, the spatial depth non-uniformity compensation device applies a point spread function (PSF) to either the patch 511 of the first image 510 or the patch 521 of the second image 520, and then includes a pre-built compensation model. May be used to determine the depth for the pixel of the edge on which the irregular comparison was performed.

The spatial depth non-uniformity compensation process using such an irregular comparison is performed before the depths of the pixels of the edges detected in the first image 510 and the second image 520 are extracted or the first image 510 and the second image ( The depth for the pixel of the edge detected at 520 may be extracted and performed at the same time.

Referring to FIG. 6, in operation 610, the spatial depth non-uniformity compensating device acquires a first image and a second image.

Subsequently, the spatial depth non-uniformity compensation device detects an edge in each of the first image and the second image (620).

The spatial depth non-uniformity compensator then compares the irregularities to the pixels of the edge based on an offset jitter map in which an offset irregular vector corresponding to each of the subregions in which the first image and the second image are divided is stored. Perform 630.

At this time, the spatial depth non-uniformity compensating apparatus sets a first patch in a preset range around the pixels of the edge detected in the first image, and the subregion including the pixels of the edge detected in the first image from the offset irregular map. After selecting an offset irregular vector corresponding to, the first highest similarity to the first patch within a preset jitter range centered on the pixel shifted by the selected offset irregular vector from the pixel of the edge detected in the second image. By obtaining two patches, one can perform an irregular comparison on the pixels at the edges.

Here, the process of acquiring the second patch may be a process of acquiring a second patch around the pixel shifted by the delta irregular vector from the selected offset irregular vector.

Thereafter, the spatial depth non-uniformity compensation device determines a depth of the pixel of the edge where the random comparison is performed, using a compensation model in which the depth according to the blur size corresponding to each of the subregions is stored (640).

At this time, the spatial depth non-uniformity compensation device extracts a depth corresponding to the blur size in the first patch and the second patch using a compensation model, and then determines the extracted depth as the depth for the pixel of the edge where the irregular comparison is performed. Can be.

In addition, although not shown in the figure, the spatial depth non-uniformity compensation device may form an offset irregular map before performing an irregular comparison with respect to the pixels of the edge.

Specifically, the spatial depth non-uniformity compensating apparatus sets a first patch of a preset range around the pixels of the edge detected in the first image, and within the preset search area around the pixels of the edge detected in the second image. After obtaining a third patch having the highest similarity with the first patch, and calculating a distance vector (jitter) between the first patch and the third patch, the distance vector calculated for the pixel of the edge included in each of the subregions By statistically setting an offset irregular vector corresponding to each of the subregions based on the offset irregular map, it is possible to form an offset irregular map.

In addition, although not shown in the figure, the spatial depth non-uniformity compensation device may build a compensation model before performing an irregular comparison with respect to pixels at the edges.

Here, the spatial depth non-uniformity compensation apparatus may construct a compensation model by setting a depth according to a blur size corresponding to each of the subregions in consideration of the center pixel and the edge pixel in each of the subregions.

Referring to FIG. 7, an apparatus for compensating spatial depth nonuniformity according to an exemplary embodiment includes an image acquirer 710, an edge detector 720, an irregular comparison performer 730, and a depth determiner 740.

The image acquirer 710 acquires a first image and a second image.

The edge detector 720 detects an edge in each of the first image and the second image.

The irregular comparison performing unit 730 performs an irregular comparison with respect to the pixels of the edge based on an offset jitter map in which an offset irregular vector corresponding to each of the subregions in which the first image and the second image are divided is stored. Perform.

In this case, the irregular comparison performing unit 730 sets a first patch having a preset range around the pixels of the edge detected in the first image, and includes the pixels of the edge detected in the first image from the offset irregular map. After selecting the offset irregular vector corresponding to the subregion, the degree of similarity with the first patch within the preset jitter range centered on the pixel shifted by the selected offset irregular vector from the pixel of the edge detected in the second image. By obtaining a high second patch, an irregular comparison can be performed on the pixels at the edges.

The depth determiner 740 determines a depth of the pixel of the edge where the random comparison is performed, using a compensation model in which the depth according to the blur size corresponding to each of the subregions is stored (640).

At this time, the depth determiner 740 extracts a depth corresponding to the blur size in the first patch and the second patch by using the compensation model, and then extracts the extracted depth as the depth of the pixel of the edge where the irregular comparison is performed. You can decide.

In addition, although not shown in the drawings, the spatial depth non-uniformity compensation device may further include an offset irregular map forming unit.

The offset irregular map forming unit may form an offset irregular map before performing an irregular comparison on the pixels of the edge.

Specifically, the offset irregular map forming unit sets a first patch in a preset range around the pixels of the edge detected in the first image, and sets the first patch within a preset search area around the pixels of the edge detected in the second image. A third patch having the highest similarity to one patch is obtained, and the distance vector jitter between the first patch and the third patch is calculated, and then the distance vector calculated for the pixel of the edge included in each of the subregions. By statistically setting the offset irregularity vector corresponding to each of the subregions, an offset irregularity map can be formed.

In addition, although not shown in the drawings, the spatial depth non-uniformity compensation device may further include a compensation model constructing unit.

The compensation model builder may build a compensation model before performing an irregular comparison on the pixels at the edges.

Here, the compensation model builder may build a compensation model by setting a depth according to a blur size corresponding to each of the subregions in consideration of the center pixel and the edge pixel in each of the subregions.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the devices and components described in the embodiments may be, for example, processors, controllers, arithmetic logic units (ALUs), digital signal processors, microcomputers, field programmable arrays (FPAs), It may be implemented using one or more general purpose or special purpose computers, such as a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of the software. For convenience of explanation, one processing device may be described as being used, but one of ordinary skill in the art will appreciate that the processing device includes a plurality of processing elements and / or a plurality of types of processing elements. It can be seen that it may include. For example, the processing device may include a plurality of processors or one processor and one controller. In addition, other processing configurations are possible, such as parallel processors.

The software may include a computer program, code, instructions, or a combination of one or more of the above, and configure the processing device to operate as desired, or process it independently or collectively. You can command the device. Software and / or data may be any type of machine, component, physical device, virtual equipment, computer storage medium or device in order to be interpreted by or to provide instructions or data to the processing device. Or may be permanently or temporarily embodied in a signal wave to be transmitted. The software may be distributed over networked computer systems so that they may be stored or executed in a distributed manner. Software and data may be stored on one or more computer readable recording media.

The method according to the embodiment may be embodied in the form of program instructions that can be executed by various computer means and recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the media may be those specially designed and constructed for the purposes of the embodiments, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, such as floppy disks. Magneto-optical media, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

Although the embodiments have been described by the limited embodiments and the drawings as described above, various modifications and variations are possible to those skilled in the art from the above description. For example, the described techniques may be performed in a different order than the described method, and / or components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different form than the described method, or other components. Or even if replaced or substituted by equivalents, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and equivalents to the claims are within the scope of the claims that follow.

Claims

In the spatial depth non-uniformity compensation method using jittered comparison,

Obtaining a first image and a second image;

Detecting an edge in each of the first image and the second image;

Performing an irregular comparison on a pixel of the edge based on an offset jitter map in which an offset irregular vector corresponding to each of the sub-regions in which the first image and the second image are divided is stored; And

Determining a depth of a pixel of the edge on which the irregular comparison is performed by using a compensation model in which a depth according to a blur size corresponding to each of the subregions is stored;

Spatial depth non-uniformity compensation method comprising a.
The method of claim 1,

Performing an irregular comparison on the pixels of the edge

Setting a first patch in a preset range around a pixel of the edge detected in the first image;

Selecting an offset irregular vector corresponding to a subregion including a pixel of the edge detected in the first image from the offset irregular map; And

Acquiring a second patch having the highest similarity to the first patch within a preset jitter range around a pixel shifted by the selected offset irregular vector from the pixel of the edge detected in the second image;

Spatial depth non-uniformity compensation method comprising a.
The method of claim 2,

Acquiring the second patch

And obtaining the second patch around the pixel shifted by the delta irregular vector from the selected offset irregular vector.
The method of claim 2,

Determining a depth for the pixel of the edge on which the irregular comparison is performed

Extracting a depth corresponding to a blur size in the first patch and the second patch using the compensation model; And

Determining the extracted depth as a depth for a pixel of the edge on which the irregular comparison is performed

Spatial depth non-uniformity compensation method comprising a.
The method of claim 1,

Forming the offset irregular map

The spatial depth non-uniformity compensation method further comprising.
The method of claim 5,

Forming the offset irregular map

Setting a first patch in a preset range around a pixel of the edge detected in the first image;

Acquiring a third patch having the highest similarity to the first patch within a preset search area around the pixels of the edge detected in the second image;

Calculating a distance vector between the first patch and the third patch; And

Statistically setting an offset irregular vector corresponding to each of the subregions based on a distance vector calculated for a pixel of an edge included in each of the subregions.

Spatial depth non-uniformity compensation method comprising a.
The method of claim 1,

Building the compensation model

The spatial depth non-uniformity compensation method further comprising.
The method of claim 7, wherein

Building the compensation model

Setting a depth according to a blur size corresponding to each of the subregions in consideration of a center pixel and an edge pixel in each of the subregions

Spatial depth non-uniformity compensation method comprising a.
A computer-readable recording medium having recorded thereon a program for performing the method of any one of claims 1 to 8.
In a spatial depth non-uniformity compensation device using jittered comparison,

An image obtaining unit obtaining the first image and the second image;

An edge detector detecting an edge in each of the first image and the second image;

Performing an irregular comparison on a pixel of the edge based on an offset jitter map storing an offset irregular vector corresponding to each of the subregions in which the first image and the second image are divided; part; And

Depth determination unit for determining the depth of the pixel of the edge on which the irregular comparison is performed using a compensation model stored in the depth according to the blur size corresponding to each of the sub-regions

Spatial depth non-uniformity compensation device comprising a.
The method of claim 10,

The irregular comparison performing unit

A first patch having a preset range is set around the pixels of the edge detected in the first image, and an offset corresponding to a subregion including the pixels of the edge detected in the first image from the offset irregular map Selecting an irregular vector, the first highest similarity to the first patch within a preset jitter range around a pixel shifted from the pixel of the edge detected in the second image by the selected offset irregular vector; A spatial depth non-uniformity compensation device that obtains two patches.
The method of claim 11,

The depth determination unit

Using the compensation model to extract a depth corresponding to the blur size in the first patch and the second patch, and determine the extracted depth as the depth for the pixel of the edge on which the random comparison was performed Depth non-uniformity compensation device.
The method of claim 10,

An offset irregular map forming unit forming the offset irregular map

Spatial depth non-uniformity compensation device further comprising.
The method of claim 13,

The offset irregular map forming unit

A first patch having a preset range is set around the pixels of the edge detected in the first image, and the first patch is located within a preset search area around the pixels of the edge detected in the second image. Obtain a third patch having the highest similarity, calculate a distance vector (jitter) between the first patch and the third patch, and based on the calculated distance vector for a pixel of an edge included in each of the subregions And statistically setting an offset irregular vector corresponding to each of the subregions.
The method of claim 10,

Compensation model building unit for building the compensation model

Spatial depth non-uniformity compensation device further comprising.
The method of claim 15,

The compensation model building unit

And setting a depth according to a blur size corresponding to each of the subregions in consideration of a center pixel and an edge pixel in each of the subregions.