KR101204210B1 - Methods of deinterlacing using geometric duality and image processing device using the same - Google Patents

Abstract

A deinterlacing method using geometric duality and an image processing apparatus for performing the method are disclosed. The deinterlacing method using geometric duplexing is based on determining the local edge direction of the interpolation target pixel and the relationship between existing transmitted pixels based on the determined local edge direction of the interpolation target pixel and the similarity between the pixels to be interpolated. And calculating the weight using the geometric duality and applying the calculated weight to the interpolation target pixel to deinterlace the interpolation target pixel. Therefore, the deinterlacing performance is improved compared to the existing deinterlacing method so that high resolution images can be received.

Description

Deinterlacing method using geometric duality and image processing device performing the method {METHODS OF DEINTERLACING USING GEOMETRIC DUALITY AND IMAGE PROCESSING DEVICE USING THE SAME}

The present invention relates to a deinterlacing method using geometric duality and an image processing apparatus for performing the method, and more particularly, to a method and apparatus for performing deinterlacing in consideration of the edge direction.

The existing SDTV broadcasting standards (NTSC, PAL SECAM) used the interlaced scan format to efficiently use a limited transmission band. Such a progressive scan method has a frame rate twice that of a progressive scan format in the same band using a field rate of 50/60 Hz that does not recognize a blink of a human eye. This means that only half of the vertical resolution of the image can be used to express high-quality vertical components in an area with little motion, and a high-field rate can represent a good image due to a high field rate. However, this catenary scanning method causes visual problems such as flickering, line tweeting, and line crawling when there is a high frequency component or a lot of movement in the vertical direction. As a result, the vertical resolution is deteriorated in the case of images having such characteristics.

In spite of these problems, the conventional scanning method has been used as a conventional TV transmission method due to the limitation of the transmission band and the technical problem, but with the recent technological development, the progressive scanning method is possible without the aforementioned problems of the conventional scanning method. Done In addition, DTV (Digital TV), which transmits high-density information, has been commercialized, and in the case of such a DTV broadcasting system, a sequential scanning method is more suitable than a conventional scan method for high quality and high definition, and in fact, a sequential scanning method is mainly adopted. In addition, sequential scanning methods are suitable for HDTV, flat panel displays (LCD, plasma TV), PC monitors, and various digital display devices, and mainly adopt sequential scanning methods. Therefore, the transition from a conventional scan method to a sequential scan method is the trend of the times and can be said to be in the transitional period.

However, in the past, most of the cultural contents have been produced by an anti-scanning method, and even in the transitional reality where the existing TV transmission system coexists, the anti-scanning method cannot be applied immediately and only the sequential scanning method can be applied. This can be a huge waste if you apply all the schemes or just the sequential scan scheme to all the systems. Therefore, an interlaced to progressive conversion (IPC) method of converting an image signal of a progressive scan method into a sequential scan method is an important problem, and this conversion process is called deinterlacing. In other words, the deinterlacing method is to interpolate the scan lines disappearing from each field of the perforated scanning method and convert them into a sequential signal form. In addition, a method of converting a scanning method is used to convert a field rate in an image signal or to output a still image of one frame in a general image. At this time, in order to effectively transmit a digital video signal, since a sequential scanning video signal is more advantageous in terms of handling and compression efficiency, a high quality high speed deinterlacing method is required.

Many deinterlacing methods have been proposed for IPC. The proposed deinterlacing methods are classified into two categories, intrafield interpolation and interfield interpolation. First, an interpolation method in a field is an interpolation method using only values of pixels in a field, which can be divided into an interpolation method using a spatial correlation such as a repetition of a row and an interpolation method using a nonlinear filter such as a media filter. . Second, interfield interpolation methods use correlations between fields, and can be classified into a motion adaptive method and a motion compensation method through motion estimation.

First, the interpolation method using spatial correlation is very simple because only pixel values of the current field are used. Therefore, these methods are widely used because of their simple hardware implementation because they do not require field memory and require low computational complexity. However, these methods are limited in finding missing pixel values by using only the information in the current field, and are difficult to obtain high quality images such as flickering and blurring in areas with high motion or high frequency components. There is this. To overcome this problem, methods for more effectively interpolating pixel values using outline information in a field have been proposed.

Second, a method of using a nonlinear filter including a median filter is also an interpolation method in a field. The median filter uses only outline information. However, if the media filter is extended to the time domain, this is implemented by switching between interfield and interfield interpolation methods similarly to the motion adaptive method. In general, the median filter has a problem in that small details in the vertical direction are distorted or cause aliasing.

Third, the motion adaptive interpolation method uses time domain information without motion estimation. Since there is a high correlation between adjacent fields, the pixel values lost in the current field are appropriately selected according to the movement type of the current field by using the information of the adjacent fields.

Finally, the motion compensation method generally performs best among various interlacing methods. These methods use the high correlation of adjacent fields to estimate the motion of the current field and compensate for the motion. If the motion estimation is wrong, unwanted pixel values can be restored, and a lot of time is spent in motion estimation and compensation. There is a problem.

In general, in-field deinterlacing methods of the above-described methods do not require a separate field memory and the time required to perform deinterlacing by simple calculation is short, but motion region and boundary region restoration are not satisfactory. On the other hand, the inter-field interlacing methods achieve many performance improvements by applying the de-interlacing method in the field, and then estimating the motion of the current field and compensating for the motion, but are still unsatisfactory in the horizontal boundary region and the high frequency region. There is a tendency to show results.

Accordingly, it is a first object of the present invention to provide a deinterlacing method using geometric duality for improving deinterlacing performance.

In addition, a second object of the present invention is to provide an image processing apparatus for performing a deinterlacing method using geometric duality for improving the deinterlacing performance.

, 중간에 위치한 화소를

, 우측에 위치한 화소를

, 상기 보간 대상 화소의 하위 스캔 라인에 위치하되 상기 보간 대상 화소를 기준으로 좌측에 위치한 화소를

, 중간에 위치한 화소를

, 우측에 위치한 화소를

라고 할 경우, 아래의 식 1에 의하여 상기 보간 대상 화소의 주변 픽셀 화소값 사이의 차이와 상기 보간 대상 화소의 주변 픽셀 화소값 사이의 공간적 코릴레이션을 판단하고, Deinterlacing method using a geometric duality according to an aspect of the present invention for achieving the first object of the present invention described above is based on the step of determining the local edge direction of the interpolation target pixel and the determined local edge direction of the interpolation target pixel Deriving a weight using geometric duplex based on a similar relationship between the existing transmitted pixels and a relationship between the pixels to be interpolated, and deinterlacing the interpolation target pixel by applying the calculated weight to the interpolation target pixel. It may include. The determining of the local edge direction of the interpolation target pixel may be performed based on three pixels included in an upper scan line of the interpolation target pixel and three pixels included in a lower scan line of the interpolation target pixel. The local edge direction can be determined. Based on the determined local edge direction of the interpolation target pixel, a weight is calculated by using geometric duality based on the similarity between the existing transmitted pixels and the relationship between the pixels to be interpolated, and the weight is calculated based on the interpolation target. The deinterlacing of the interpolation target pixel by applying to a pixel is performed based on three pixels included in an upper scan line of the interpolation target pixel and three pixels included in a lower scan line of the interpolation target pixel. A weight to be applied to the interpolation target pixel may be calculated using pixel values of pixels having a local edge direction. The determining of the local edge direction of the interpolation target pixel may include a pixel located on an upper scan line of the interpolation target pixel and positioned on the left side of the interpolation target pixel.

, The middle pixel

, The pixel on the right

A pixel positioned on a lower scan line of the interpolation target pixel and positioned on the left side of the interpolation target pixel

, The middle pixel

, The pixel on the right

In this case, according to Equation 1 below, the spatial correlation between the difference between the pixel value of the pixel of the interpolation target pixel and the pixel value of the pixel of the pixel of the interpolation target pixel is determined.

(Equation 1)

은 상기

과 상기

의 화소값 차의 절대값, 상기

는 상기

와 상기

의 화소값 차의 절대값,

은 상기

와 상기

의 화소값 차의 절대값이고

는 상기

과

,

와

,

와

의 화소값 차의 절대값을 더해 3으로 나눈 값,

는

과

,

와

의 화소값 차의 절대값을 더해 2로 나눈 값,

는

와

,

와

의 화소값 차의 절대값을 더해 2로 나눈 값을 의미함) 상기 식 1를 이용해 산출된 값을 기초로,

인 경우, 상기 보간 대상 화소의 상기 로컬 에지 방향을

방향-여기서, 상기

방향은 보간 대상 화소를 중심으로 한 스캔 라인 상단 좌측에 위치한 화소 및 한 스캔 라인 수직 아래 위치한 화소를 연결한 방향임-으로 판별하고,Where

Said above

And

Absolute value of the pixel value difference of

Above

And

Absolute value of the difference in pixel values,

Said above

And

Is the absolute value of the pixel value difference

Above

and

,

Wow

,

Wow

The absolute value of the difference in pixel values divided by 3

The

and

,

Wow

Plus the absolute value of the difference in pixel values divided by 2,

The

Wow

,

Wow

Means the absolute value of the difference in pixel values divided by 2.) Based on the value calculated using Equation 1,

When is, the local edge direction of the interpolation target pixel

Direction-where

The direction is determined by connecting the pixel located at the upper left of the scan line centered on the interpolation target pixel and the pixel positioned vertically below one scan line.

인 경우 상기 상기 보간 대상 화소의 로컬 에지 방향을

방향-여기서, 상기

방향은 보간 대상 화소를 중심으로 한 스캔 라인 수직 상단에 위치한 화소 및 한 스캔 라인 하단 좌측에 위치한 화소를 연결한 방향임-으로 판별하고,

In the case of the local edge direction of the interpolation target pixel.

Direction-where

The direction is determined by connecting the pixel located at the top of the scan line vertically centered on the interpolation target pixel and the pixel located at the bottom left of one scan line.

In other cases, the local edge direction may be determined as the vertical direction of the interpolation target pixel. Based on the determined local edge direction of the interpolation target pixel, a weight is calculated by using geometric duality based on a similar relationship between pixels to be interpolated, and the calculated weight is interpolated. Deinterlacing the interpolation target pixel by applying the target pixel,

방향-여기서, 상기

방향은 보간 대상 화소를 중심으로 한 스캔 라인 상단 좌측에 위치한 화소 및 한 스캔 라인 수직 아래 위치한 화소를 연결한 방향임-으로 판단된 경우 하기 식 2The local edge direction of the interpolation target pixel

Direction-where

Direction is determined by connecting the pixel located at the upper left of the scan line centered on the interpolation target pixel and the pixel located below the vertical line of one scan line.

(Equation 2)

는 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소-여기서, 상기 3개의 화소는 보간 대상 화소의 수직 윗방향, 좌측 상단 방향, 우측 상단 방향에 위치한 화소임-와 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소-여기서, 상기 3개의 화소는 보간 대상 화소의 수직 아래 방향, 좌측 하단 방향, 우측 하단 방향에 위치한 화소임-로 이루어진 행렬이고, 상기

행렬의 각 행은 보간 대상 화소가 i번째 행 j번째 열에 위치한 경우,

에 포함된 각 화소는 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소는 i-1번째 행에 j-1, j, j+1열에 위치한 화소 및 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소는 i+1번째 행에 j-1, j, j+1열에 위치한 화소이고,

행렬의 각 행에 포함된 화소는 상기 상위 스캔 라인에 포함된 3개의 화소 및 상기 하위 스캔 라인에 포함된 3개의 화소 중 하나의 화소 중심으로 상기 하나의 화소가 i-1번째 행에 j-1번째 열에 위치한 화소인 경우, 두 스캔 라인 수직 위에 위치한 화소-여기서, 상기 두 스캔 라인 수직 위에 위치한 화소는 i-3번째 행 및 j-1번째 열에 위치한 화소임-, 두 스캔 라인 수직 아래에 위치한 화소- 여기서 상기 두 스캔 라인 수직 아래에 위치한 화소는 i+1번째 행에 j-1번째 열에 위치한 화소임-, 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소- 여기서, 상기 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소는 i-3번째 행에 j-3번째 열에 위치한 화소임-, 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소-여기서, 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소는 i+1번째 행에 j+1번째 열에 위치한 화소임-고,

는

로서 상기

행렬과 상기

행렬을 기초로 산출된 가중치 벡터로 기하학적 이중성을 기초로 산출된 상기 보간 대상 화소의 화소값을 보간하는데 사용되는 가중치 벡터임)(

Are three pixels included in the upper scan line of the interpolation target pixel, wherein the three pixels are pixels located in the vertical upward direction, upper left direction, and upper right direction of the interpolation target pixel, and lower scan lines of the interpolation target pixel. Is a matrix consisting of three pixels included in the pixel, wherein the three pixels are pixels positioned in a vertically downward direction, a lower left direction, and a lower right direction of the interpolation target pixel.

For each row of the matrix, if the pixel to be interpolated is located in the i th row j th column,

Each pixel included in is 3 pixels included in the upper scan line of the interpolation target pixel, and the 3 pixels included in the j-1, j, j + 1 columns in the i-1 th row and 3 included in the lower scan line of the interpolation target pixel. Pixels are pixels j-1, j, j + 1 in the i + 1th row,

Pixels included in each row of the matrix are centered on one pixel among three pixels included in the upper scan line and three pixels included in the lower scan line. The pixel located in the first column, the pixels located vertically above the two scan lines, wherein the pixel located above the two scan lines vertically are the pixels located in the i-3th row and the j-1 th column; Wherein the pixels located below the two scan lines vertically are the pixels located in the j-1th column of the i + 1th row, the pixels located above the two scan lines vertically and two pixels to the left. Pixels located above and to the left of two pixels are pixels located in column j-3 of the i-3th row, below two scan lines vertically, and pixels located next to two pixels to the right, where two scan lines vertically below Ooh The pixel located next to the two pixels on the i + 1-th row j + 1-th column in the pixel being - and,

The

As above

Matrix and above

Is a weight vector calculated based on a matrix and is a weight vector used to interpolate pixel values of the pixel to be interpolated based on geometric duality.

It may include the step of calculating the weight using. Based on the determined local edge direction of the interpolation target pixel, a weight is calculated by using geometric duality based on a similar relationship between pixels to be interpolated, and the calculated weight is interpolated. Deinterlacing the interpolation target pixel by applying the target pixel,

를 기초로 아래의 식 3Calculated above

Based on Equation 3 below

(Equation 3)

The method may include interpolating and deinterlacing the interpolation target pixel using the lateral interpolation. Based on the determined local edge direction of the interpolation target pixel, a weight is calculated by using geometric duality based on a similar relationship between pixels to be interpolated, and the calculated weight is interpolated. Deinterlacing the interpolation target pixel by applying the target pixel,

방향-여기서, 상기

방향은 보간 대상 화소를 중심으로 한 스캔 라인 수직 상단에 위치한 화소 및 한 스캔 라인 하단 좌측에 위치한 화소를 연결한 방향-으로 판단된 경우 아래의 식 4The local edge direction of the interpolation target pixel

Direction-where

Direction is determined by connecting the pixel located at the vertical top of the scan line centered on the interpolation target pixel and the pixel located at the bottom left of one scan line.

(Equation 4)

는 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소(보간 대상 화소의 수직 윗방향, 좌측 상단 방향, 우측 상단 방향에 위치한 화소)와 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소(보간 대상 화소의 수직 아래 방향, 좌측 하단 방향, 우측 하단 방향)로 이루어진 행렬,

행렬의 각 행은 보간 대상 화소가 i번째 행 j번째 열에 위치한 경우,

에 포함된 각 화소는 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소는 i-1번째 행에 j-1, j, j+1열에 위치한 화소 및 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소는 i+1번째 행에 j-1, j, j+1열에 위치한 화소이고,

행렬의 각 행에 포함된 화소는 상기 상위 스캔 라인에 포함된 3개의 화소 및 상기 하위 스캔 라인에 포함된 3개의 화소 중 하나의 화소 중심으로 상기 하나의 화소가 i-1번째 행에 j-1번째 열에 위치한 화소인 경우, 두 스캔 라인 수직 위에 위치한 화소-여기서, 상기 두 스캔 라인 수직 위에 위치한 화소는 i-3번째 행에 j-1번째 열에 위치한 화소임-, 두 스캔 라인 수직 아래에 위치한 화소-여기서, 상기 두 스캔 라인 수직 아래에 위치한 화소는 i+1번째 행에 j-1번째 열에 위치한 화소임-, 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소-여기서, 상기 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소는 i-3번째 행에 j+1번째 열에 위치한 화소임-, 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소-여기서, 상기 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소는 i+1번째 행에 j-3번째 열에 위치한 화소임-고,

는

로서 상기

행렬과 상기

행렬을 기초로 산출된 가중치 벡터로 기하학적 이중성을 이용하여 상기 보간 대상 화소의 화소값을 보간하는데 사용되는 가중치 벡터임.)(

Are three pixels included in the upper scan line of the interpolation target pixel (pixels located in the vertical upward direction, upper left direction, and upper right direction of the interpolation target pixel) and three pixels included in the lower scan line of the interpolation target pixel (interpolation). Matrix consisting of a vertically downward direction, a lower left direction, and a lower right direction of the target pixel;

For each row of the matrix, if the pixel to be interpolated is located in the i th row j th column,

Each pixel included in is 3 pixels included in the upper scan line of the interpolation target pixel, and the 3 pixels included in the j-1, j, j + 1 columns in the i-1 th row and 3 included in the lower scan line of the interpolation target pixel. Pixels are pixels j-1, j, j + 1 in the i + 1th row,

Pixels included in each row of the matrix are centered on one pixel among three pixels included in the upper scan line and three pixels included in the lower scan line. The pixel located in the first column, the pixel located above the two scan lines vertically, wherein the pixel located above the two scan line vertically are the pixels located in the j-1 th column in the i-3 th row, Wherein the pixels located below the two scan lines vertically are the pixels located in the j-1th column of the i + 1th row, and the pixels located above the two scan lines vertically and two pixels to the left, where the two scan lines Pixels located above two pixels to the left and above vertical are pixels located in the j + 1th column in the i-3th row, two pixels below the scan line vertical and two pixels to the right, where the two scan lines vertical bottom, The pixels next to the two pixels to the right are the pixels located in the j-3th column of the i + 1th row.

The

As above

Matrix and above

A weight vector calculated based on a matrix, which is a weight vector used to interpolate pixel values of the pixel to be interpolated using geometric duality.)

It may include the step of calculating the weight using. Based on the determined local edge direction of the interpolation target pixel, a weight is calculated by using geometric duality based on a similar relationship between pixels to be interpolated, and the calculated weight is interpolated. Deinterlacing the interpolation target pixel by applying the target pixel,

를 기초로 아래의 식 5 Calculated above

Based on Equation 5 below

(Eq. 5)

The method may include interpolating and deinterlacing the interpolation target pixel using the lateral interpolation. Based on the determined local edge direction of the interpolation target pixel, a weight is calculated by using geometric duality based on a similar relationship between pixels to be interpolated, and the calculated weight is interpolated. Deinterlacing the interpolation target pixel by applying the target pixel,

방향 또는

방향을 제외한 방향으로 판단되는 경우 아래의 식 6 The local edge direction of the interpolation target pixel

Direction or

Equation 6 below when it is judged as a direction except

(Equation 6)

Deinterlacing the interpolation target pixel by using a.

In addition, the image processing apparatus for performing the deinterlacing using geometric duality according to an aspect of the present invention for achieving the second object of the present invention described above is a local edge of the interpolation target pixel based on the peripheral pixel information of the received interpolation target pixel. A local edge direction determination unit for determining a direction, and based on a local edge direction of the interpolation target pixel determined by the local edge direction determination unit, geometric duality based on a similar relationship between the pixels to be interpolated The apparatus may include a weight calculator configured to calculate a weight to be applied to the interpolation target pixel, and a pixel interpolator to apply the weight of the interpolation target pixel calculated through the weight calculator to the interpolation target pixel. The local edge direction determination unit may determine a local edge direction of the interpolation target pixel based on three pixels included in an upper scan line of the interpolation target pixel and three pixels included in a lower scan line of the interpolation target pixel. . The weight calculator includes pixel values of pixels having the same local edge direction as the interpolation target pixel based on three pixels included in the upper scan line of the interpolation target pixel and three pixels included in the lower scan line of the interpolation target pixel. A weight to be applied to the interpolation target pixel may be calculated using. The local edge direction determination unit,

, 중간에 위치한 화소를

, 우측에 위치한 화소를

, 상기 보간 대상 화소의 하위 스캔 라인에 위치하되 상기 보간 대상 화소를 기준으로 좌측에 위치한 화소를

, 중간에 위치한 화소를

, 우측에 위치한 화소를

라고 할 경우, 아래의 식 1에 의하여 상기 보간 대상 화소의 주변 픽셀 화소값 사이의 차이와 상기 보간 대상 화소의 주변 픽셀 화소값 사이의 공간적 코릴레이션을 판단하고, Pixels positioned on the upper scan line of the interpolation target pixel and positioned on the left side of the interpolation target pixel

, The middle pixel

, The pixel on the right

A pixel positioned on a lower scan line of the interpolation target pixel and positioned on the left side of the interpolation target pixel

, The middle pixel

, The pixel on the right

In this case, according to Equation 1 below, the spatial correlation between the difference between the pixel value of the pixel of the interpolation target pixel and the pixel value of the pixel of the pixel of the interpolation target pixel is determined.

(Equation 1)

은 상기

과 상기

의 화소값 차의 절대값, 상기

는 상기

와 상기

의 화소값 차의 절대값,

은 상기

와 상기

의 화소값 차의 절대값이고

는 상기

과

,

와

,

와

의 화소값 차의 절대값을 더해 3으로 나눈 값,

는

과

,

와

의 화소값 차의 절대값을 더해 2로 나눈 값,

는

와

,

와

의 화소값 차의 절대값을 더해 2로 나눈 값을 의미한다.)Where

Said above

And

Absolute value of the pixel value difference of

Above

And

Absolute value of the difference in pixel values,

Said above

And

Is the absolute value of the pixel value difference

Above

and

,

Wow

,

Wow

The absolute value of the difference in pixel values divided by 3

The

and

,

Wow

Plus the absolute value of the difference in pixel values divided by 2,

The

Wow

,

Wow

Means the absolute value of the pixel value difference, divided by 2.

Based on the value calculated using Equation 1,

인 경우, 상기 로컬 에지 방향을

방향-여기서, 상기

방향은 보간 대상 화소를 중심으로 한 스캔 라인 상단 좌측에 위치한 화소 및 한 스캔 라인 수직 아래 위치한 화소를 연결한 방향임-으로 판별하고,

If is, the local edge direction

Direction-where

The direction is determined by connecting the pixel located at the upper left of the scan line centered on the interpolation target pixel and the pixel positioned vertically below one scan line.

인 경우 상기 로컬 에지 방향을

방향-여기서, 상기

방향은 보간 대상 화소를 중심으로 한 스캔 라인 수직 상단에 위치한 화소 및 한 스캔 라인 하단 좌측에 위치한 화소를 연결한 방향임-으로 판별하고,

If is the local edge direction

Direction-where

The direction is determined by connecting the pixel located at the top of the scan line vertically centered on the interpolation target pixel and the pixel located at the bottom left of one scan line.

In other cases, the local edge direction may be determined as the vertical direction of the interpolation target pixel. The weight calculation unit,

방향-여기서, 상기

방향은 보간 대상 화소를 중심으로 한 스캔 라인 상단 좌측에 위치한 화소 및 한 스캔 라인 수직 아래 위치한 화소를 연결한 방향임-으로 판단된 경우 아래의 식 2The local edge direction of the interpolation target pixel is changed through the local edge direction determination unit.

Direction-where

The direction is the direction connecting the pixel located on the upper left of the scan line centered on the interpolation target pixel and the pixel located below the vertical line of one scan line.

(Equation 2)

는 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소-여기서, 상기 3개의 화소는 보간 대상 화소의 수직 윗방향, 좌측 상단 방향, 우측 상단 방향에 위치한 화소임-와 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소-여기서, 상기 3개의 화소는 보간 대상 화소의 수직 아래 방향, 좌측 하단 방향, 우측 하단 방향에 위치한 화소임-로 이루어진 행렬이고, 상기

행렬의 각 행은 보간 대상 화소가 i번째 행 j번째 열에 위치한 경우,

에 포함된 각 화소는 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소는 i-1번째 행에 j-1, j, j+1열에 위치한 화소 및 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소는 i+1번째 행에 j-1, j, j+1열에 위치한 화소이고,

행렬의 각 행에 포함된 화소는 상기 상위 스캔 라인에 포함된 3개의 화소 및 상기 하위 스캔 라인에 포함된 3개의 화소 중 하나의 화소 중심으로 상기 하나의 화소가 i-1번째 행에 j-1번째 열에 위치한 화소인 경우, 두 스캔 라인 수직 위에 위치한 화소-여기서, 상기 두 스캔 라인 수직 위에 위치한 화소는 i-3번째 행 및 j-1번째 열에 위치한 화소임-, 두 스캔 라인 수직 아래에 위치한 화소- 여기서 상기 두 스캔 라인 수직 아래에 위치한 화소는 i+1번째 행에 j-1번째 열에 위치한 화소임-, 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소- 여기서, 상기 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소는 i-3번째 행에 j-3번째 열에 위치한 화소임-, 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소-여기서, 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소는 i+1번째 행에 j+1번째 열에 위치한 화소임-고,

는

로서 상기

행렬과 상기

행렬을 기초로 산출된 가중치 벡터로 기하학적 이중성을 기초로 산출된 상기 보간 대상 화소의 화소값을 보간하는데 사용되는 가중치 벡터임)를 이용하여 상기 가중치를 산출하는 단계를 포함할 수 있다. 상기 화소 보간부는,(

Are three pixels included in the upper scan line of the interpolation target pixel, wherein the three pixels are pixels located in the vertical upward direction, upper left direction, and upper right direction of the interpolation target pixel, and lower scan lines of the interpolation target pixel. Is a matrix consisting of three pixels included in the pixel, wherein the three pixels are pixels positioned in a vertically downward direction, a lower left direction, and a lower right direction of the interpolation target pixel.

For each row of the matrix, if the pixel to be interpolated is located in the i th row j th column,

Each pixel included in is 3 pixels included in the upper scan line of the interpolation target pixel, and the 3 pixels included in the j-1, j, j + 1 columns in the i-1 th row and 3 included in the lower scan line of the interpolation target pixel. Pixels are pixels j-1, j, j + 1 in the i + 1th row,

Pixels included in each row of the matrix are centered on one pixel among three pixels included in the upper scan line and three pixels included in the lower scan line. The pixel located in the first column, the pixels located vertically above the two scan lines, wherein the pixel located above the two scan lines vertically are the pixels located in the i-3th row and the j-1 th column; Wherein the pixels located below the two scan lines vertically are the pixels located in the j-1th column of the i + 1th row, the pixels located above the two scan lines vertically and two pixels to the left. Pixels located above and to the left of two pixels are pixels located in column j-3 of the i-3th row, below two scan lines vertically, and pixels located next to two pixels to the right, where two scan lines vertically below Ooh The pixel located next to the two pixels on the i + 1-th row j + 1-th column in the pixel being - and,

The

As above

Matrix and above

And calculating the weight using a weight vector calculated based on a matrix, which is a weight vector used to interpolate pixel values of the pixel to be interpolated based on geometric duality. The pixel interpolation unit,

를 기초로 아래의 식 3The calculated by the local edge discriminating unit

Based on Equation 3 below

(Equation 3)

The method may include interpolating and deinterlacing the interpolation target pixel using the lateral interpolation. The weight calculation unit,

방향-여기서, 상기

방향은 보간 대상 화소를 중심으로 한 스캔 라인 수직 상단에 위치한 화소 및 한 스캔 라인 하단 좌측에 위치한 화소를 연결한 방향-으로 판단된 경우, 아래의 식 4The local edge direction of the interpolation target pixel is changed through the local edge direction determination unit.

Direction-where

The direction is determined by connecting the pixel located at the vertical top of the scan line centered on the interpolation target pixel and the pixel located at the bottom left of one scan line.

(Equation 4)

는 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소(보간 대상 화소의 수직 윗방향, 좌측 상단 방향, 우측 상단 방향에 위치한 화소)와 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소(보간 대상 화소의 수직 아래 방향, 좌측 하단 방향, 우측 하단 방향)로 이루어진 행렬,

행렬의 각 행은 보간 대상 화소가 i번째 행 j번째 열에 위치한 경우,

에 포함된 각 화소는 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소는 i-1번째 행에 j-1, j, j+1열에 위치한 화소 및 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소는 i+1번째 행에 j-1, j, j+1열에 위치한 화소이고,

행렬의 각 행에 포함된 화소는 상기 상위 스캔 라인에 포함된 3개의 화소 및 상기 하위 스캔 라인에 포함된 3개의 화소 중 하나의 화소 중심으로 상기 하나의 화소가 i-1번째 행에 j-1번째 열에 위치한 화소인 경우, 두 스캔 라인 수직 위에 위치한 화소-여기서, 상기 두 스캔 라인 수직 위에 위치한 화소는 i-3번째 행에 j-1번째 열에 위치한 화소임-, 두 스캔 라인 수직 아래에 위치한 화소-여기서, 상기 두 스캔 라인 수직 아래에 위치한 화소는 i+1번째 행에 j-1번째 열에 위치한 화소임-, 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소-여기서, 상기 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소는 i-3번째 행에 j+1번째 열에 위치한 화소임-, 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소-여기서, 상기 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소는 i+1번째 행에 j-3번째 열에 위치한 화소임-고,

는

로서 상기

행렬과 상기

행렬을 기초로 산출된 가중치 벡터로 기하학적 이중성을 이용하여 상기 보간 대상 화소의 화소값을 보간하는데 사용되는 가중치 벡터임.)(

Are three pixels included in the upper scan line of the interpolation target pixel (pixels located in the vertical upward direction, upper left direction, and upper right direction of the interpolation target pixel) and three pixels included in the lower scan line of the interpolation target pixel (interpolation). Matrix consisting of a vertically downward direction, a lower left direction, and a lower right direction of the target pixel;

For each row of the matrix, if the pixel to be interpolated is located in the i th row j th column,

Each pixel included in is 3 pixels included in the upper scan line of the interpolation target pixel, and the 3 pixels included in the j-1, j, j + 1 columns in the i-1 th row and 3 included in the lower scan line of the interpolation target pixel. Pixels are pixels j-1, j, j + 1 in the i + 1th row,

Pixels included in each row of the matrix are centered on one pixel among three pixels included in the upper scan line and three pixels included in the lower scan line. The pixel located in the first column, the pixel located above the two scan lines vertically, wherein the pixel located above the two scan line vertically are the pixels located in the j-1 th column in the i-3 th row, Wherein the pixels located below the two scan lines vertically are the pixels located in the j-1th column of the i + 1th row, and the pixels located above the two scan lines vertically and two pixels to the left, where the two scan lines Pixels located above two pixels to the left and above vertical are pixels located in the j + 1th column in the i-3th row, two pixels below the scan line vertical and two pixels to the right, where the two scan lines vertical bottom, The pixels next to the two pixels to the right are the pixels located in the j-3th column of the i + 1th row.

The

As above

Matrix and above

A weight vector calculated based on a matrix, which is a weight vector used to interpolate pixel values of the pixel to be interpolated using geometric duality.)

It may include the step of calculating the weight using. The pixel interpolation unit,

를 기초로 아래의 식 5The calculated by the local edge discriminating unit

Based on Equation 5 below

(Eq. 5)

The method may include interpolating and deinterlacing the interpolation target pixel using. The pixel interpolation unit,

방향 또는

방향을 제외한 방향으로 판단되는 경우 아래의 식 6The local edge direction of the interpolation target pixel is changed through the local edge direction determination unit.

Direction or

Equation 6 below when it is judged as a direction except

(Equation 6)

And deinterlacing the interpolation target pixel using. An image processing apparatus that performs deinterlacing using the geometric duplex may be used in a high definition television (HDTV) receiver or a TV receiver having a resolution greater than or equal to HDTV.

As described above, according to the deinterlacing method using geometric duplex according to the embodiment of the present invention and the image processing apparatus for performing the method, the deinterlacing is performed by applying individual weights to peripheral pixels of the interpolation target pixel using the geometric duplex. .

Therefore, the deinterlacing performance is improved over the conventional deinterlacing method so that high resolution images can be received.

In addition, when the size of the display is increased, it is possible to obtain a satisfactory picture quality while enabling real-time processing while reducing processing complexity as much as possible when used in a TV receiver of HDTV or higher.

방향인 경우, 보간 대상 화소를 복원하는 방법을 나타내는 개념도이다.
도 4는 본 발명의 일실시예에 따른 로컬 에지 방향이

방향인 경우, 보간 대상 화소를 복원하는 방법을 나타내는 개념도이다.
도 5는 본 발명의 일실시예에 따른 디인터레이싱 방법을 기존의 디인터레이싱 방법과 비교한 표이다.
도 6은 본 발명의 일실시예에 따른 기하학적 이중성을 이용한 디인터레이싱 방법을 사용하여 디인터레이싱을 수행한 영상의 성능을 나타내기 위한 것이다.
도 7은 본 발명의 일실시예에 따른 기하학적 이중성을 이용하여 디인터레이싱 방법을 나타내는 순서도이다.
도 8은 본 발명의 일실시예에 따른 기하학적 이중성을 이용하여 디인터레이싱을 수행하는 영상 처리 장치를 나타낸 개념도이다. 1 is a conceptual diagram illustrating a CAD algorithm for deinterlacing.
FIG. 2 is a conceptual diagram illustrating a modified edge-based line averaging (MELA) technique among deinterlacing methods.
3 is a local edge direction according to an embodiment of the present invention

In the case of a direction, it is a conceptual diagram which shows the method of restoring an interpolation target pixel.
4 is a local edge direction according to an embodiment of the present invention

In the case of a direction, it is a conceptual diagram which shows the method of restoring an interpolation target pixel.
5 is a table comparing a deinterlacing method with a conventional deinterlacing method according to an embodiment of the present invention.
FIG. 6 illustrates the performance of an image subjected to deinterlacing using a deinterlacing method using geometric duality according to an embodiment of the present invention.
7 is a flowchart illustrating a deinterlacing method using geometric duality according to an embodiment of the present invention.
8 is a conceptual diagram illustrating an image processing apparatus that performs deinterlacing using geometric duality according to an embodiment of the present invention.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. Like reference numerals are used for like elements in describing each drawing.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there is no other component in between.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Hereinafter, the same reference numerals are used for the same components in the drawings, and duplicate descriptions of the same components are omitted.

Hereinafter, although the embodiment of the present invention discloses a method of interpolating assuming a window size of 1, the window size for deinterlacing may have a value of 1 or more, as long as it does not emerge from the essence of the present invention.

In addition, according to an embodiment of the present invention, the pixel value used in the present invention may mean a luminance value of the pixel for convenience of description. However, the color difference value of the pixel may also be included in the scope of the present invention as a concept included in the pixel value, unless it deviates from the essence of the present invention.

1 is a conceptual diagram illustrating a CAD algorithm for deinterlacing.

)에 대한 8 방향의 기준 가중치 계수

가 표시되어 있으며, 아래의 수학식 1로부터 가중치 계수

를 구할 수 있다.Referring to Figure 1, the reference pixel (

Reference weighting factor in 8 directions for

Is shown, and the weighting coefficient is

Can be obtained.

는 인터레이스 이미지(즉, 저해상도 이미지)에서 로컬 윈From here,

Is a local window in an interlaced image (ie, a low resolution image).

It means eight pixel values adjacent to m, which is a target pixel for position interpolation, in a local window.

)에 대한 8개의 에지 방향의 기준 가중치는

이고, 보간 대상 화소에 대한 8개의 에지 방향의 적용 가중치는

이다. 일반적인 자연 영상은 인접한 화소의 값이 크게 변하지 않는 특성이 있다. 그러므로, 에지의 방향도 일정 범위 내에서 동일한 경향성을 가진다. 그리하여, 두 변수 기준 가중치

, 적용 가중치

의 상관도가 양의 상관 관계를 나타내며, 두 변수

,

의 공분산의 값도 양수가 된다. 이는 아래의 수학식 2로 표현된다.Referring to FIG. 1, a specific reference pixel (

The reference weights for the eight edge directions for

The applied weights of eight edge directions for the interpolation target pixel are

to be. The general natural image has a characteristic that the value of adjacent pixels does not change significantly. Therefore, the direction of the edge also has the same tendency within a certain range. Thus, two variable reference weights

, Weight applied

The correlation of represents a positive correlation, and the two variables

,

The covariance of is also positive. This is represented by Equation 2 below.

는 기준 화소의 각 방향에 대한 기준 가중치 벡터,

는 보상 대상 화소의 각 방향에 대한 적용 가중치 벡터이다. 즉, 기준 화소로부터 얻어지는 기준 가중치를 이용하여 보간 대상 화소에 적용되는 적용 가중치를 구할 수 있다.From here,

Is a reference weight vector for each direction of the reference pixel,

Is an applied weight vector for each direction of the compensation target pixel. That is, the application weight applied to the interpolation target pixel may be obtained using the reference weight obtained from the reference pixel.

In CAD, interpolation target pixels are interpolated by using geometric duality that the relationship between the transmitted pixels is the same as the relationship between the pixels to be interpolated. That is, it is assumed that the relationship between the previously transmitted pixels is the same as the relationship between the pixels to be interpolated. Such a relationship may be referred to as geometric duality.

Accordingly, pixels having the same local edge direction based on geometric duality may have the same weight or have a weight in proportional relationship.

FIG. 2 is a conceptual diagram illustrating a modified edge-based line averaging (MELA) technique among deinterlacing methods.

Referring to FIG. 2, the MELA technique may interpolate pixel values using three pixels located above and three pixels located below the interpolation target pixel.

는 보간 대상 화소의 이웃 화소들의 값을 나타낸다. 보간 대상 화소의 상위 스캔 라인에 위치하되 보간 대상 화소를 기준으로 좌측에 위치한 화소를

, 중간에 위치한 화소를

, 우측에 위치한 화소를

, 보간 대상 화소의 하위 스캔 라인에 위치하되 보간 대상 화소를 기준으로 좌측에 위치한 화소를

, 중간에 위치한 화소를

, 우측에 위치한 화소를

라고 할 수 있다.

Denotes values of neighboring pixels of the interpolation target pixel. Pixels positioned on the upper scan line of the interpolation target pixel but left on the interpolation target pixel

, The middle pixel

, The pixel on the right

, Located on the lower scan line of the interpolation target pixel, but located on the left side of the pixel

, The middle pixel

, The pixel on the right

.

The MELA technique uses the Local Edge Direction using three Directional Pixel Difference Measures represented by Equation 3 below and Directional Correlation represented by Equation 4 below. ) Can be determined.

는

번째 줄,

번째 열에 위치한 보간되는 화소값을 의미한다.

기호는 AND 연산을 나타낸다. 즉, MELA 기법에서는

값과

값을 모두 고려하여 화소를 보간한다. 수학식 5에 표현된 상단의 두 개의 수식을 보면, 보간 대상 화소값의 로컬 에지 방향이 대각선 방향으로 선택되면 그 방향의 선상에 놓여있는 인접한 화소 4개의 평균값을 보간값으로 예측한다. 그 외의 경우, 수학식 5 하단의 수학식과 같이 보간하고자 하는 화소의 위와 아래의 화소값의 평균값을 보간값으로 결정한다. here,

The

Line one,

The interpolated pixel value located in the first column.

The symbol represents an AND operation. In other words, in the MELA technique

Value and

The pixels are interpolated considering all the values. Referring to the two equations at the top of Equation 5, when the local edge direction of the interpolation target pixel value is selected in the diagonal direction, the average value of four adjacent pixels lying on the line in the direction is predicted as the interpolation value. In other cases, as shown in Equation 5 below, the average value of pixel values above and below the pixel to be interpolated is determined as the interpolation value.

방향은 보간 대상 화소를 중심으로 한 스캔 라인 상단 좌측에 위치한 화소 및 한 스캔 라인 수직 아래 위치한 화소를 연결한 방향이고,

방향은 보간 대상 화소를 중심으로 한 스캔 라인 수직 상단에 위치한 화소 및 한 스캔 라인 하단 좌측에 위치한 화소를 연결한 방향이다.

The direction is a direction connecting pixels located on the upper left side of the scan line centered on the interpolation target pixel and pixels positioned vertically below one scan line.

The direction is a direction connecting pixels located on the upper vertical line of the scan line centered on the interpolation target pixel and pixels located on the lower left side of one scan line.

The local edge direction of the interpolation target pixel may be determined based on three pixels included in the upper scan line of the interpolation target pixel and three pixels included in the lower scan line of the interpolation target pixel.

방향으로 판단하는 경우, 보간 대상 화소를 중심으로 수직 윗방향으로

, 수직 아래 방향으로

, 좌측 상단 방향으로

, 우측 하단 방향

에 위치한 화소의 화소값을 이용하여 보간 대상화소를 보간한다.That is, the local edge direction

In the case of judging in the direction, it is vertically upward with respect to the interpolation target pixel.

, Vertically downward

, In the upper left direction

, Bottom right direction

The interpolation target pixel is interpolated using the pixel value of the pixel located at.

방향으로 판단하는 경우, 보간 대상 화소를 중심으로 수직 윗방향으로

, 수직 아래 방향으로

, 우측 상단 방향으로

, 좌측 하단 방향

에 위치한 화소의 화소값을 이용하여 보간 대상 화소를 보간한다.Also, the local edge orientation

In the case of judging in the direction, it is vertically upward with respect to the interpolation target pixel.

, Vertically downward

, In the upper right direction

, Bottom left direction

The interpolation target pixel is interpolated using the pixel value of the pixel located at.

, 수직 아래 방향으로

에 위치한 화소의 화소값을 이용하여 보간 대상 화소를 보간한다.
In addition, when it is determined to be in the local edge direction, it is vertically upward with respect to the interpolation target pixel.

, Vertically downward

The interpolation target pixel is interpolated using the pixel value of the pixel located at.

방향으로 판단하는 경우, 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소와 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소 중 하나의 화소를 기준으로 수직 방향으로 위로 두 스캔 라인 위에 있는 화소에

, 수직 방향으로 두 스캔 라인 아래에 있는 화소에

, 수직 방향 위로 두 스캔 라인 위 좌측 두 화소 옆에 위치한 화소에

, 수직 방햐으로 두 스캔 라인 아래 우측 두 화소 옆에 위치한 화소에

의 가중치를 부여하여 이용하여 보간 대상 화소를 보간할 수 있다.
In the deinterlacing method using geometric duality according to an embodiment of the present invention, the local edge direction is determined.

In the case of judging in the direction, the pixels above the two scan lines in the vertical direction with respect to one of the three pixels included in the upper scan line of the interpolation target pixel and one of the three pixels included in the lower scan line of the interpolation target pixel. on

On the pixels below the two scan lines in the vertical direction

On the pixels next to the two left pixels above the two scan lines

On the pixel located next to the right two pixels below the two scan lines in a vertical direction.

The interpolation target pixel may be interpolated by using a weighting factor of.

According to an embodiment of the present invention, the present invention may interpolate an interpolation target pixel by using Equation 6, which is a modification of the MELA technique.

인 경우

,

,

,

각각의 화소값에 가중치

,

,

,

를 곱하여 더해주고,

인 경우,

,

,

,

각각의 화소값에 가중치

,

,

,

를 곱하여 더해준다. 그 외의 경우,

와

의 화소값의 평균을 보간 대상 화소의 화소값으로 할 수 있다.
Referring to Equation 6,

If

,

,

,

Weight each pixel value

,

,

,

Multiply by to add it,

Quot;

,

,

,

Weight each pixel value

,

,

,

Multiply and add. Otherwise,

Wow

The average of pixel values can be taken as the pixel value of the interpolation target pixel.

방향인 경우, 보간 대상 화소를 복원하는 방법을 나타내는 개념도이다. 3 is a local edge direction according to an embodiment of the present invention

In the case of a direction, it is a conceptual diagram which shows the method of restoring an interpolation target pixel.

Referring to FIG. 3, the conceptual diagram includes six pixels D10, D11, D12, D17, D18, and D19 that exist around the interpolation target pixel and the interpolation target pixel, and pixels D1 that exist around six pixels. , D2, D3, D4, D5, D6, D7, D8, D9, D13, D14, D15, D16, D20, D21, D22, D23, D24, D25, D26, D27, D28).

라고

행렬의 각 행은 각 데이터 벡터의 원소의

방향에 놓여있는 4개의 이웃한 화소들이다. Referring to equation (7), the pixel around the interpolation target pixel data vector

Called

Each row of the matrix is an element of each data vector

Four neighboring pixels in the direction.

방향으로 판단하는 경우, 보간 대상 화소를 중심으로 수직으로 한 스캔 라인 윗 방향에 위치한 화소에

, 수직으로 한 스캔 라인 아래 방향에 위치한 화소에

, 수직으로 한 스캔 라인 위, 좌측으로 하나 옆에 위치한 화소에

, 수직으로 한 스캔 라인 아래 우측 하나 옆에 위치한 화소에

의 가중치를 부여하여 보간 대상화소를 보간할 수 있다. In the conventional MELA method, the local edge direction of the target pixel is changed.

Direction, the pixel located above the scan line perpendicular to the interpolation target pixel.

To the pixels vertically below one scan line

On the pixels vertically on one scan line and one on the left

On the pixels located next to the right one below the scan line

An interpolation target pixel may be interpolated by assigning a weight to.

방향으로 판단하는 경우, 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소(보간 대상 화소의 수직 윗방향, 좌측 상단 방향, 우측 상단 방향에 위치한 화소)와 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소(보간 대상 화소의 수직 아래 방향, 좌측 하단 방향, 우측 하단 방향) 중 하나의 화소를 기준으로 산출된 수직 방향으로 두 스캔 라인 위에 위치한 화소에서

, 수직 방향으로 아래로 두 스캔 라인 아래에 위치한 화소에서

, 수직 방향으로 두 스캔 라인 위 좌측으로 두 화소 옆에 위치한 화소에서

, 수직 방향으로 두 스캔 라인 아래 우측으로 두 화소 옆에 위치한 화소에서

의 가중치를 산출하고, 기하학적 이중성을 기초로 동일한 가중치를 보간 대상 화소를 보간하기 위한 가중치로 산출할 수 있다. In the deinterlacing method using geometric duality according to an embodiment of the present invention, the local edge direction is determined.

In the case of judging in the direction, three pixels included in the upper scan line of the interpolation target pixel (pixels located in the vertical upward direction, upper left direction, and upper right direction of the interpolation target pixel) and the lower scan line of the interpolation target pixel In a pixel located above two scan lines in a vertical direction calculated based on one of three pixels (vertical downward direction, lower left direction, and lower right direction of the interpolation target pixel).

From the pixels located below the two scan lines down in the vertical direction,

At pixels next to two pixels to the left above both scan lines in the vertical direction

At pixels next to two pixels to the right below two scan lines in the vertical direction,

The weight of may be calculated and the same weight may be calculated as the weight for interpolating the interpolation target pixel based on the geometric duplex.

That is, when the interpolation target pixel is located in the i-th row of the j-th column, the three pixels included in the upper scan line of the interpolation target pixel mean the pixels located in the j-1, j, j + 1 columns of the i-1 th row. The three pixels included in the lower scan line of the interpolation target pixel mean pixels located in j-1, j, and j + 1 columns in the i + 1th row.

방향으로 판단하는 경우, i-1번째 행에 j-1번째 열에 위치한 화소를 기준으로 두 스캔 라인 수직 위에 위치한 화소( i-3번째 행에 j-1번째 열에 위치한 화소), 두 스캔 라인 수직 아래에 위치한 화소( i+1번째 행에 j-1번째 열에 위치한 화소), 두 스캔 라인 수직 위, 좌측으로 두 화소 옆에 위치한 화소( i-3번째 행에 j-3번째 열에 위치한 화소), 두 스캔 라인 수직 아래, 우측으로 두 화소 옆에 위치한 화소( i+1번째 행에 j+1번째 열에 위치한 화소)를 기초로 보간 대상 화소에 적용할 가중치를 산출할 수 있다. According to one embodiment of the invention the local edge direction is

In the case of judging in the direction, the pixels located above the two scan lines vertically based on the pixels located in the j-1th column in the i-1th row (the pixels located in the j-1th row in the i-3th row), and the two vertically below the scanline. Pixels located in the j-1th column of the i + 1th row, pixels located above the two scan lines vertically and two pixels to the left (pixels located in the j-3th column of the i-3th row), two A weight to be applied to the interpolation target pixel may be calculated based on the pixels (pixels located in the j + 1th column of the i + 1th row) positioned two pixels below the scan line vertically and to the right.

행렬의 첫 번째 행을 참조하면,

행렬의 첫 번째 행에는 D1, D3, D17, D19 의 화소값이 행렬에 표현된다. 개념도를 보면, D1, D3, D17, D19는 보간 대상 화소의 주변 화소 중 하나인 D10을 기준으로

방향의 로컬 에지 방향에 놓여있는 4개의 화소값이다. E.g,

If you refer to the first row of the matrix,

In the first row of the matrix, pixel values of D1, D3, D17, and D19 are represented in the matrix. In the conceptual diagram, D1, D3, D17, and D19 are based on D10, which is one of the pixels around the interpolation target pixel.

Four pixel values lying in the local edge direction of the direction.

행렬은 각 행은 순차적으로 보간 대상 화소의 주변에 존재하는 6개의 화소(D10, D11, D12, D17, D18, D19)를 기준으로

방향의 로컬 에지 방향에 놓여있는 4개의 화소값으로 구성될 수 있다.

The matrix is based on six pixels (D10, D11, D12, D17, D18, and D19) that exist in the periphery of the pixel to be interpolated sequentially.

It may consist of four pixel values lying in the local edge direction of the direction.

Equation 8 below shows the weight application according to the geometric duality according to an embodiment of the present invention.

방향의 로컬 에지 방향에 놓여있는 4개의 화소값을 이용하여 산출한 가중치 계수인

는 기하학적 이중성에 의해 보간 대상 화소를 기준으로

방향의 로컬 에지 방향에 놓인 주변 화소에 적용하는 가중치

와 유사한 값을 가질 수 있다. 따라서, 기존이 알고 있는 화소값을 이용하여

를 산출함으로써 보간 대상 화소를 기준으로

방향의 로컬 에지 방향에 놓인 주변 화소에 적용하는 가중치

를 산출할 수 있다. That is, based on the reference pixel (three pixels included in the upper scan line of the interpolation target pixel and three pixels included in the lower scan line of the interpolation target pixel)

Weight coefficient calculated using four pixel values lying in the local edge direction of the

Is based on the interpolation target pixel by geometric duality

Weight applied to surrounding pixels oriented in the local edge direction of the direction

It can have a value similar to Therefore, by using the known pixel value

Is calculated based on the interpolation target pixel.

Weight applied to surrounding pixels oriented in the local edge direction of the direction

Can be calculated.

That is, interpolation is performed using pixel values of pixels having the same local edge direction as the interpolation target pixel based on three pixels included in the upper scan line of the interpolation target pixel and three pixels included in the lower scan line of the interpolation target pixel. The weight to be applied to the target pixel may be calculated.

와 데이터 행렬

는 아래의 수학식 9와 같은 관계를 가질 수 있다. Data Vector of Equation 9

And data matrix

May have a relationship as in Equation 9 below.

는 최소 자승 근사값(Least Square Approximation)을 이용하여 아래의 수학식 10을 사용하여 구할 수 있다.
In Equation 9 above

Can be obtained using Equation 10 below using Least Square Approximation.

값을 유추하여 보간 대상 화소를 보간할 수 있다.
That is, using geometric duality through the weight value calculated using Equation 10

The interpolation target pixel may be interpolated by inferring a value.

방향인 경우, 보간 대상 화소를 복원하는 방법을 나타내는 개념도이다. 4 is a local edge direction according to an embodiment of the present invention

In the case of a direction, it is a conceptual diagram which shows the method of restoring an interpolation target pixel.

Referring to FIG. 4, the conceptual diagram includes six pixels D10, D11, D12, D17, D18, and D19 that exist around the interpolation target pixel and the interpolation target pixel, and pixels D1 that exist around the six pixels. , D2, D3, D4, D5, D6, D7, D8, D9, D13, D14, D15, D16, D20, D21, D22, D23, D24, D25, D26, D27, D28).

라고

행렬의 각 행은 각 데이터 벡터의 원소의

방향에 놓여있는 4개의 이웃한 화소들이다. Referring to Equation 11, the pixel of the peripheral of the interpolation target pixel is a data vector

Called

Each row of the matrix is an element of each data vector

Four neighboring pixels in the direction.

방향으로 판단하는 경우, 보간 대상 화소를 중심으로 수직 윗 방향에 위치한 화소에

, 수직 아랫 방향에 위치한 화소에

, 수직 윗방향, 우측 한 화소 옆에 위치한 화소에

, 수직 아래 방향, 좌측 한 화소 옆에 위치한 화소에

의 가중치를 부여하여 보간 대상화소를 보간할 수 있다.In the conventional MELA deinterlacing method, the local edge direction of the pixel to be interpolated

In the case of judging in the direction, the pixel located in the vertical upward direction with respect to the interpolation target pixel

On the pixels vertically down

, Vertically upward, to the pixel next to one pixel on the right

, Vertically down, to the pixel next to one pixel to the left

An interpolation target pixel may be interpolated by assigning a weight to.

방향으로 판단하는 경우, 보간 대상 화소의 상위 스캔 라인에 포함된 3개의 화소(보간 대상 화소의 수직 윗방향, 좌측 상단 방향, 우측 상단 방향에 위치한 화소)와 보간 대상 화소의 하위 스캔 라인에 포함된 3개의 화소(보간 대상 화소의 수직 아래 방향, 좌측 하단 방향, 우측 하단 방향) 중 하나의 화소를 기준으로 수직 방향으로 두 스캔 라인 위에 위치한 화소에서

, 수직 방향으로 두 스캔 라인 아래에 위치한 화소에서

, 수직으로 두 스캔 라인 위 우측으로 두 화소 옆에 위치한 화소에서

, 수직으로 두 스캔 라인 위 우측으로 두 화소 옆에 위치한 화소에서

의 가중치를 산출하고 기하학적 이중성을 기초로 이를 보간 대상 화소를 보간하기 위해 적용할 수 있다.In the deinterlacing method using geometric duality according to an embodiment of the present invention, the local edge direction is determined.

In the case of judging in the direction, three pixels included in the upper scan line of the interpolation target pixel (pixels located in the vertical upward direction, upper left direction, and upper right direction of the interpolation target pixel) and the lower scan line of the interpolation target pixel Pixels located above two scan lines in a vertical direction with respect to one of three pixels (vertical downward direction, lower left direction, and lower right direction of the interpolation target pixel)

In pixels located below two scan lines in the vertical direction,

At pixels next to two pixels vertically above the two scan lines

At pixels next to two pixels vertically above the two scan lines

Calculate the weight of and apply it to interpolate the interpolation target pixel based on the geometric duality.

That is, when the interpolation target pixel is located in the i-th row of the j-th column, the three pixels included in the upper scan line of the interpolation target pixel mean pixels located in the j-1, j, j + 1 columns of the i-1 th row, respectively. The three pixels included in the lower scan line of the interpolation target pixel mean pixels located in j-1, j, and j + 1 columns in the i + 1th row, respectively.

방향으로 판단하는 경우, i-1번째 행에 j-1번째 열에 위치한 화소를 기준으로 두 스캔 라인 수직 위에 위치한 화소( i-3번째 행에 j-1번째 열에 위치한 화소), 두 스캔 라인 수직 아래에 위치한 화소( i+1번째 행에 j-1번째 열에 위치한 화소), 두 스캔 라인 수직 위, 우측으로 두 화소 옆에 위치한 화소( i-3번째 행에 j+1번째 열에 위치한 화소), 두 스캔 라인 수직 아래, 좌측으로 두 화소 옆에 위치한 화소( i+1번째 행에 j-3번째 열에 위치한 화소)를 기초로 보간 대상 화소에 적용할 가중치를 산출할 수 있다.
According to one embodiment of the invention the local edge direction is

In the case of judging in the direction, the pixels located above the two scan lines vertically based on the pixels located in the j-1th column in the i-1th row (the pixels located in the j-1th row in the i-3th row), and the two vertically below the scanline. Pixels located in the j-1th column of the i + 1th row, pixels located above the two scan lines vertically and two pixels to the right (pixels located in the j + 1th column of the i-3th row), two A weight to be applied to the interpolation target pixel may be calculated based on the pixels (pixels located in the j-3th column of the i + 1th row) positioned two pixels below the scan line vertically and to the left.

행렬의 첫 번째 행을 참조하면,

행렬의 첫 번째 행에는 D3, D5, D15, D17 의 화소값이 행렬에 표현된다. 개념도를 보면, D3, D5, D15, D17은 보간 대상 화소의 주변 화소 중 하나인 D10을 기준으로

방향의 로컬 에지 방향에 놓여있는 4개의 화소값이다. E.g,

If you refer to the first row of the matrix,

In the first row of the matrix, pixel values of D3, D5, D15, and D17 are represented in the matrix. In the conceptual diagram, D3, D5, D15, and D17 are based on D10, which is one of the surrounding pixels of the interpolation target pixel.

Four pixel values lying in the local edge direction of the direction.

행렬은 각 행은 순차적으로 보간 대상 화소의 주변에 존재하는 6개의 화소(D10, D11, D12, D17, D18, D19)를 기준으로

방향의 로컬 에지 방향에 놓여있는 4개의 화소값으로 구성될 수 있다.

The matrix is based on six pixels (D10, D11, D12, D17, D18, and D19) that exist in the periphery of the pixel to be interpolated sequentially.

It may consist of four pixel values lying in the local edge direction of the direction.

Equation 12 below shows a weight application according to geometric duality according to an embodiment of the present invention.

방향의 로컬 에지 방향에 놓여있는 4개의 화소값을 이용하여 산출한 가중치 계수인

는 기하학적 이중성에 의해 보간 대상 화소를 기준으로

방향의 로컬 에지 방향에 놓인 주변 화소에 적용하는 가중치

와 유사한 값을 가질 수 있다. 따라서, 기존이 알고 있는 화소값을 이용하여

를 산출함으로써 보간 대상 화소를 기준으로

방향의 로컬 에지 방향에 놓인 주변 화소에 적용하는 가중치

를 산출할 수 있다. That is, based on the reference pixel (three pixels included in the upper scan line of the interpolation target pixel and three pixels included in the lower scan line of the interpolation target pixel)

Weight coefficient calculated using four pixel values lying in the local edge direction of the

Is based on the interpolation target pixel by geometric duality

Weight applied to surrounding pixels oriented in the local edge direction of the direction

It can have a value similar to Therefore, by using the known pixel value

Is calculated based on the interpolation target pixel.

Weight applied to surrounding pixels oriented in the local edge direction of the direction

Can be calculated.

That is, interpolation is performed using pixel values of pixels having the same local edge direction as the interpolation target pixel based on three pixels included in the upper scan line of the interpolation target pixel and three pixels included in the lower scan line of the interpolation target pixel. The weight to be applied to the target pixel may be calculated.

와 데이터 행렬

는 아래의 수학식 13과 같은 관계를 가질 수 있다. Data Vector of Equation 11

And data matrix

May have a relationship as shown in Equation 13 below.

는 최소 자승 근사값(Least Square Approximation)을 이용하여 아래의 수학식 14를 사용하여 구할 수 있다. In Equation 13 above

Can be obtained using Equation 14 below using Least Square Approximation.

값을 유추하는 방법으로 보간 대상 화소를 보간할 수 있다.
That is, by using the geometric duality through the weight value calculated using Equation 14

The interpolation target pixel may be interpolated by inferring a value.

The CAD method described in FIG. 1 has excellent performance, but has 20 times more processing time than the conventional deinterlacing method using LA. Moreover, the eight neighbors for calculating correlation between Low-Resolution Samples (previously decoded and known pixel values) because the CAD method does not consider local edge orientation. Pixels may be included unnecessarily.

According to one embodiment of the present invention, the method considering the geometric duality disclosed in the present invention may consider only fewer pixels in the local edge direction than the CAD method, in order to overcome the disadvantage of the CAD method.

That is, according to an embodiment of the present invention, the deinterlacing method considering the geometric duality may have a better deinterlacing performance while the processing speed is much improved than the conventional CAD method in consideration of pixel values in the local edge direction.

5 is a table comparing a deinterlacing method with a conventional deinterlacing method according to an embodiment of the present invention.

영상들 (Airplane, Boat, Girl), 3개의

영상(Football, Foreman, Table, Tennis), 하나의

영상(Raven), 하나의

영상(Blue Sky)에 대해 실험을 한 것이다.Referring to FIG. 5, three deinterlacing methods for geometric duality according to an embodiment of the present invention include three conventional deinterlacing techniques (LA, ELA, EELA, DOI, NEDD, MELA, FDD, EPD, and CAD).

Videos (Airplane, Boat, Girl), 3

Video (Football, Foreman, Table, Tennis), one

Raven, one

We experimented with the image (Blue Sky).

높고 CAD 방법보다 0.5

높다. Referring to the table of FIG. 5, the deinterlacing method using geometric duplex according to the present invention has an average of 0.4 than the MELA method in PSNR.

0.5 higher than CAD method

high.

In particular, it can be seen that the deinterlacing method using the geometric duality according to the present invention has a good reconstruction performance for an image having a diagonal edge like the Foreman image.

In addition, the deinterlacing method using the geometrical duality according to the present invention has a processing speed that is reduced by more than 79% compared to the CAD method in a high resolution image.

FIG. 6 illustrates the performance of an image subjected to deinterlacing using a deinterlacing method using geometric duality according to an embodiment of the present invention.

Referring to FIG. 6, each screen compares the subjective image quality of the first frame of the Raven sequence.

LA, ELA, EELA, DOI, NEDD, MELA, FDD, EPD, CAD, the deinterlacing method according to the invention in alphabetical order from the top left.

The LA and FDD methods provide relatively high PSNR results, but in detail the graininess and staircase artifacts appear in the image.

The deinterlacing method using NEDD, MELA, CAD and the geometric duplex according to the present invention has satisfactory results compared to the LA and FDD methods.

Although the CAD method preserves the edges of the original image well, blurring occurs in the background area. In contrast, the deinterlacing method using geometric duality according to the present invention has an image almost similar to the original image. ELAs, EELAs, DOIs, and EPDs can identify noise caused by inaccurate detection in the local edge direction.

 7 is a flowchart illustrating a deinterlacing method using geometric duality according to an embodiment of the present invention.

Referring to FIG. 7, the local edge direction of the interpolation target pixel may be calculated (S700).

The local edge direction of the interpolation target pixel is determined through the conditional expression of Equation 6 based on the values calculated through Equations 3 and 4 using the MELA method.

인 경우, 로컬 에지 방향을

방향으로 결정하고

인 경우, 로컬 에지 방향을

방향으로 결정하고 그 이외의 경우 로컬 에지 방향을 보간 대상 화소의 수직 방향으로 결정한다.

If, the local edge direction

Decided in the direction

If, the local edge direction

Direction, and otherwise, the local edge direction is determined as the vertical direction of the interpolation target pixel.

The weight is calculated based on the determined local edge direction (step S710).

방향으로 결정된 경우, 수학식 7 내지 수학식 10을 이용하여 가중치

를 산출한다. 보간 대상 화소의 로컬 에지 방향을

방향으로 결정된 경우, 수학식 11 내지 수학식 14를 이용하여 가중치

를 산출한다. Direction of the local edge of the pixel

When determined in the direction, the weight using the equations (7) to (10)

Calculate Direction of the local edge of the pixel

When determined in the direction, weights using Equations 11 to 14

Calculate

The interpolation target pixel is deinterlaced by applying the calculated weight (step S720).

방향으로 결정된 경우,

,

,

,

각각의 화소값에 가중치

, ,

,

를 곱하여 더해주고, 보간 대상 화소의 로컬 에지 방향이

방향으로 결정된 경우,

,

,

,

각각의 화소값에 가중치

,

,

,

를 곱하여 더해준다. 그 외의 경우,

와

의 화소값의 평균을 보간 대상 화소의 화소값으로 할 수 있다.
By using Equation 6, the local edge direction of the interpolation target pixel

If determined by direction,

,

,

,

Weight each pixel value

, ,

,

Multiply by to get the local edge direction of the pixel

If determined by direction,

,

,

,

Weight each pixel value

,

,

,

Multiply and add. Otherwise,

Wow

The average of pixel values can be taken as the pixel value of the interpolation target pixel.

8 is a conceptual diagram illustrating an image processing apparatus that performs deinterlacing using geometric duality according to an embodiment of the present invention.

For convenience of explanation, the respective components are arranged in the respective components, and at least two of the components may be combined to form one component, or one component may be divided into a plurality of components to perform the functions. The scope of the present invention is also encompassed within the scope of the present invention unless otherwise specified.

Referring to FIG. 8, an image processing apparatus that performs deinterlacing using geometric duality may include a local edge direction determination unit 800, a weight calculator 810, and a pixel interpolator 820.

The local edge direction determination unit 800 determines a local edge direction of the interpolation target pixel based on the received peripheral pixel information of the interpolation target pixel.

The local edge direction of the interpolation target pixel is determined through the conditional expression of Equation 6 based on the values calculated through Equations 3 and 4 using the MELA method.

인 경우, 로컬 에지 방향을

방향으로 결정하고

인 경우, 로컬 에지 방향을

방향으로 결정하고 그 이외의 경우 로컬 에지 방향을 보간 대상 화소의 수직 방향으로 결정한다. In other words,

If, the local edge direction

Decided in the direction

If, the local edge direction

Direction, and otherwise, the local edge direction is determined as the vertical direction of the interpolation target pixel.

The weight calculator 810 calculates a weight to be applied to the interpolation target pixel based on the local edge direction determined by the local edge direction determiner 800.

방향으로 결정된 경우, 수학식 7 내지 수학식 10을 이용하여 가중치

를 산출한다. 보간 대상 화소의 로컬 에지 방향을

방향으로 결정된 경우, 수학식 11 내지 수학식 14를 이용하여 가중치

를 산출한다. The local edge direction of the pixel to be interpolated

When determined in the direction, the weight using the equations (7) to (10)

Calculate Direction of the local edge of the pixel

When determined in the direction, weights using Equations 11 to 14

Calculate

The pixel interpolator 820 applies the weights of the interpolation target pixels calculated by the weight calculator 810 to the interpolation target pixels.

방향으로 결정된 경우,

,

,

,

각각의 화소값에 가중치

,

,

,

를 곱하여 더해주고, 보간 대상 화소의 로컬 에지 방향이

방향으로 결정된 경우,

,

,

,

각각의 화소값에 가중치

,

,

,

를 곱하여 더해준다. 그 외의 경우,

와

의 화소값의 평균을 보간 대상 화소의 화소값으로 할 수 있다.
By using Equation 4, the local edge direction of the interpolation target pixel

If determined by direction,

,

,

,

Weight each pixel value

,

,

,

Multiply by to get the local edge direction of the pixel

If determined by direction,

,

,

,

Weight each pixel value

,

,

,

Multiply and add. Otherwise,

Wow

The average of pixel values can be taken as the pixel value of the interpolation target pixel.

An image processing apparatus that performs deinterlacing according to an embodiment of the present invention may be used in an HDTV receiver. In particular, an image processing apparatus that performs deinterlacing according to an embodiment of the present invention may be used when post-processing at an HDTV receiver. In addition, the image processing apparatus for performing deinterlacing according to an embodiment of the present invention may be used when post-processing at a TV receiver having a resolution of HDTV or higher.

Although described with reference to the embodiments above, those skilled in the art will understand that the present invention can be variously modified and changed without departing from the spirit and scope of the invention as set forth in the claims below. Could be.

Claims (19)

In the deinterlacing method using geometric duplex,
Determining a local edge direction of the interpolation target pixel; And
Selecting a pixel having the same local edge direction as the local edge direction of the interpolation target pixel, calculating a weight between the pixels based on the selected pixel, and applying the calculated weight to the interpolation target pixel to deinterlace the interpolation target pixel A method of deinterlacing using geometrical duality comprising steps. The method of claim 1, wherein the determining of the local edge direction of the interpolation target pixel comprises:
The geometric duality of the interpolation target pixel may be determined based on three pixels included in an upper scan line of the interpolation target pixel and three pixels included in a lower scan line of the interpolation target pixel. Deinterlacing method used. delete The method of claim 1, wherein the determining of the local edge direction of the interpolation target pixel comprises:
Pixels positioned on the upper scan line of the interpolation target pixel and positioned on the left side of the interpolation target pixel

, The middle pixel

, The pixel on the right

A pixel positioned on a lower scan line of the interpolation target pixel and positioned on the left side of the interpolation target pixel

, The middle pixel

, The pixel on the right

In this case, according to Equation 1 below, the spatial correlation between the difference between the pixel value of the pixel of the interpolation target pixel and the pixel value of the pixel of the pixel of the interpolation target pixel is determined.
(Equation 1)

Where

Said above

And

Absolute value of the pixel value difference of

Above

And

Absolute value of the difference in pixel values,

Said above

And

Is the absolute value of the pixel value difference

Above

and

,

Wow

,

Wow

The absolute value of the difference in pixel values divided by 3

The

and

,

Wow

Plus the absolute value of the difference in pixel values divided by 2,

The

Wow

,

Wow

Means the absolute value of the difference in pixel values divided by 2)
Based on the value calculated using Equation 1,

When is, the local edge direction of the interpolation target pixel

Direction-where

The direction is determined by connecting the pixel located at the upper left of the scan line centered on the interpolation target pixel and the pixel positioned vertically below one scan line.

In the case of the local edge direction of the interpolation target pixel.

Direction-where

The direction is determined by connecting the pixel located at the top of the scan line vertically centered on the interpolation target pixel and the pixel located at the bottom left of one scan line.
In other cases, the local edge direction is determined as the vertical direction of the interpolation target pixel. The method of claim 1, wherein the deinterlacing the interpolation target pixel by applying the calculated weight to the interpolation target pixel comprises:
The local edge direction of the interpolation target pixel

Direction-where

Direction is determined by connecting the pixel located at the upper left of the scan line centered on the interpolation target pixel and the pixel located below the vertical line of one scan line.
(Equation 2)

(

Are three pixels included in the upper scan line of the interpolation target pixel, wherein the three pixels are pixels located in the vertical upward direction, upper left direction, and upper right direction of the interpolation target pixel, and lower scan lines of the interpolation target pixel. Is a matrix consisting of three pixels included in the pixel, wherein the three pixels are pixels positioned in a vertically downward direction, a lower left direction, and a lower right direction of the interpolation target pixel.

For each row of the matrix, if the pixel to be interpolated is located in the i th row j th column,

Each pixel included in is 3 pixels included in the upper scan line of the interpolation target pixel, and the 3 pixels included in the j-1, j, j + 1 columns in the i-1 th row and 3 included in the lower scan line of the interpolation target pixel. Pixels are pixels j-1, j, j + 1 in the i + 1th row,

Pixels included in each row of the matrix are centered on one pixel among three pixels included in the upper scan line and three pixels included in the lower scan line. The pixel located in the first column, the pixels located vertically above the two scan lines, wherein the pixel located above the two scan lines vertically are the pixels located in the i-3th row and the j-1 th column; Wherein the pixels located below the two scan lines vertically are the pixels located in the j-1th column of the i + 1th row, the pixels located above the two scan lines vertically and two pixels to the left. Pixels located above and to the left of two pixels are pixels located in column j-3 of the i-3th row, below two scan lines vertically, and pixels located next to two pixels to the right, where two scan lines vertically below Ooh The pixel located next to the two pixels on the i + 1-th row j + 1-th column in the pixel being - and,

The

As above

Matrix and above

Is a weight vector calculated based on a matrix and is a weight vector used to interpolate pixel values of the pixel to be interpolated based on geometric duality.
Deinterlacing method using geometric duality comprising the step of calculating the weight using. The method of claim 5, wherein the deinterlacing the interpolation target pixel by applying the calculated weight to the interpolation target pixel comprises:
Calculated above

Based on Equation 3 below
(Equation 3)

Deinterlacing method using geometric duality comprising the step of interpolating and interpolating the interpolation target pixel using. The method of claim 1, wherein the deinterlacing the interpolation target pixel by applying the calculated weight to the interpolation target pixel comprises:
The local edge direction of the interpolation target pixel

Direction-where

Direction is determined by connecting the pixel located at the vertical top of the scan line centered on the interpolation target pixel and the pixel located at the bottom left of one scan line.
(Equation 4)

(

Are three pixels included in the upper scan line of the interpolation target pixel (pixels located in the vertical upward direction, upper left direction, and upper right direction of the interpolation target pixel) and three pixels included in the lower scan line of the interpolation target pixel (interpolation). Matrix consisting of a vertically downward direction, a lower left direction, and a lower right direction of the target pixel;

For each row of the matrix, if the pixel to be interpolated is located in the i th row j th column,

Each pixel included in is 3 pixels included in the upper scan line of the interpolation target pixel, and the 3 pixels included in the j-1, j, j + 1 columns in the i-1 th row and 3 included in the lower scan line of the interpolation target pixel. Pixels are pixels j-1, j, j + 1 in the i + 1th row,

Pixels included in each row of the matrix are centered on one pixel among three pixels included in the upper scan line and three pixels included in the lower scan line. The pixel located in the first column, the pixel located above the two scan lines vertically, wherein the pixel located above the two scan line vertically are the pixels located in the j-1 th column in the i-3 th row, Wherein the pixels located below the two scan lines vertically are the pixels located in the j-1th column of the i + 1th row, and the pixels located above the two scan lines vertically and two pixels to the left, where the two scan lines Pixels located above two pixels to the left and above vertical are pixels located in the j + 1th column in the i-3th row, two pixels below the scan line vertical and two pixels to the right, where the two scan lines vertical bottom, The pixels next to the two pixels to the right are the pixels located in the j-3th column of the i + 1th row.

The

As above

Matrix and above

A weight vector calculated based on a matrix, which is a weight vector used to interpolate pixel values of the pixel to be interpolated using geometric duality.)
Deinterlacing method using the geometric duality comprising the step of calculating the weight using. The method of claim 7, wherein the deinterlacing the interpolation target pixel by applying the calculated weight to the interpolation target pixel comprises:
Calculated above

Based on Equation 5 below
(Equation 5)

Deinterlacing method using geometric duality comprising interpolating and deinterlacing the interpolation target pixel using. The method of claim 1, wherein the deinterlacing the interpolation target pixel by applying the calculated weight to the interpolation target pixel comprises:
The local edge direction of the interpolation target pixel

Direction or

Equation 6 below when it is judged as a direction except
(Equation 6)

A method of deinterlacing using geometrical duality, the method comprising: deinterlacing an interpolation target pixel using the interpolation target pixel. An image processing apparatus for performing deinterlacing using geometric duality,
A local edge direction determination unit determining a local edge direction of the interpolation target pixel based on the received peripheral pixel information of the interpolation target pixel;
A weight calculator configured to select a pixel having the same local edge direction as the local edge direction of the interpolation target pixel, and calculate a weight between pixels based on the selected pixel; And
And an interpolation unit including a pixel interpolation unit applying the weight calculated by the weight calculator to the interpolation target pixel. The method of claim 10, wherein the local edge direction determination unit,
Deinterlacing using geometric duplexing to determine the local edge direction of the interpolation target pixel based on three pixels included in the upper scan line of the interpolation target pixel and three pixels included in the lower scan line of the interpolation target pixel Image processing device. delete delete delete delete delete delete delete delete

Images