KR20060096203A - Bad edit detection method of film mode detection method - Google Patents

Abstract

The present invention relates to an editing error detection method of the film mode detection method, and in the related art, there is a problem that severe image quality deterioration occurs when the edit mode of the image source is not detected. In consideration of such a problem, the present invention includes the steps of: calculating a weighted field difference between front and rear fields based on a predetermined field of an image source; Determining a position of a predetermined field in the frame and performing a case comparison between the weighted field differences of the front and back fields according to the position to detect an editing error; When the editing error is not detected, it compares the ratio of the weight field difference of the front and rear field with the threshold and detects the last editing error.It prevents the deterioration of image quality in the film mode by accurately detecting the editing screen in the progressive image conversion of the film mode. It is effective.

Description

Editing error detection method of film mode detection method {BAD EDIT DETECTION METHOD OF FILM MODE DETECTION METHOD}

1 is an exemplary view showing an operation of converting an NTSC image into a progressive image in a conventional 3: 2 pull-down method.

2 is a block diagram showing the configuration of an image processing apparatus for explaining the present invention.

3 is a flowchart illustrating an operation of calculating an accumulated weight pixel value of a pixel difference as a criterion of an editing error in a film mode detection method according to an exemplary embodiment of the present invention.

4 is an exemplary view illustrating an operation of detecting an editing error of an edited video during an operation of converting the edited video into a progressive video according to an exemplary embodiment of the present invention.

The present invention relates to an editing error detection method of the film mode detection method. In particular, when the film mode is broken when converting the interlaced video into a progressive video, the interlaced video is switched to the video mode without performing 3: 2 pulldown. The present invention relates to an editing error detection method of the film mode detection method which enables to convert a to a progressive image.

The movie, the National Television Standards Committee (NTSC), has a different frame structure, and the movie image consists of 24 frames per second, and the NTSC image consists of 60 odd / even fields, 30 frames per second. Since the movie image is composed of 24 frames per second and the NTSC image is composed of 60 fields, when converting the movie image to NTSC image, the encoder converts one frame of the film image, which is the movie image, into three fields of the NTSC image, and the next film image. The frame is divided into two fields of the NTSC video for conversion.

FIG. 1 is an exemplary diagram illustrating an operation of converting an NTSC image into a progressive image in a conventional 3: 2 pulldown method. The image processing apparatus converts the interlaced image of 60 Hz into a progressive image of 60 Hz by 3: 2 pull-down. However, if it is edited in fields if4 and if5 and converted to another image or edited in field if 3, the image processing apparatus should convert from field of2 to progressive video using another method without performing 3: 2 pulldown.

The editing error detection method detects an image editing part when the film mode is cut off due to the image editing in the film mode, switches from the film mode to the video mode, and prevents an incorrect 3: 2 pulldown. If the image processing unit does not detect the image editing part, severe image quality deterioration occurs during the film mode.

Therefore, the present invention was devised in view of the above problems, and when the interlaced video is converted to a progressive video, the film mode is detected and the video editing portion is detected to perform the video mode without performing 3: 2 pulldown. It is an object of the present invention to provide an editing error detection method of the film mode detection method which converts and converts an interlaced image into a progressive image.

The present invention for achieving the above object, the step of calculating the weight field difference of the front and back field on the basis of the predetermined field of the image source; Determining a position of a predetermined field in the frame and performing a case comparison between the weighted field differences of the front and back fields according to the position to detect an editing error; When the editing error is not detected, it is characterized in that the step of detecting the last editing error by comparing the ratio of the weight field difference of the front and rear fields and the threshold.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 2 is a block diagram showing a configuration of an image processing apparatus for explaining the present invention, and includes an image source input unit 21 for inputting an image source as shown therein; A field difference calculator 22 for accumulating the weighted pixel difference for each pixel in the field; An editing error detection unit 23 for detecting an editing error; The image converting unit 24 does not perform 3: 2 pulldown in the film mode when detecting an editing error.

The image source input unit 21 inputs an image source, and the field difference calculator 22 calculates the difference between the front and rear fields of the image source, and calculates the weighted pixel difference by increasing the weight in a portion where the difference between the pixel difference values is large. The weighted pixel difference for each pixel in the field is accumulated.

The editing error detection unit 23 performs the editing error detection operation differently according to the position of the field at the center of the field difference to detect the editing error and informs the image converting unit 24 of the detection of the editing error.

When the editing error is detected, the image converting unit 24 converts from the film mode to the video mode and converts the progressive image using another method without performing 3: 2 pulldown.

3 is a flowchart illustrating an operation of calculating a cumulative weighted pixel value of a pixel difference that is a criterion for editing error in a film mode detection method according to an exemplary embodiment of the present invention. As shown in FIG. Calculating a difference between the pixels of the front and back fields; Calculating the weighted pixel difference by further comparing the pixel difference values by increasing the weights in proportion to the difference between the pixel difference values; Accumulating the weighted pixel difference for each pixel in the field.

The image processing apparatus obtains a difference between pixels of the front and back fields based on one field. When the image processing apparatus calculates the difference between the pixels of the front and back fields in the virtual code,

diff1 (i) = abs (f1 (i)-f2 '(i))

diff2 (i) = abs (f2 '(i)-f3 (i)), where f1 (i) is the first field pixel, f2' (i) is the second field average pixel, and f3 (i) Is the third field pixel, diff1 (i) is the first pixel difference, and diff2 (i) is the second pixel difference. In this case, f2 '(i) is a new pixel made by averaging between the field line and the line.

The image processing apparatus compares the level of the pixel differences, calculates the weighted pixel difference by increasing the weight in a portion where the difference in the pixel difference is large, and accumulates the weighted pixel difference for each pixel in the field. When the operation of accumulating the weighted pixel difference in the image processing apparatus is expressed as virtual code,

if (diff1 (i) <128)

 acc1 + = diff1 (i);

else if (diff1 (i)> 128 && diff1 (i) <128 + 32)

 acc1 + = diff1 (i) * 2;

else if (diff1 (i)> 128 + 32 && diff1 (i) <128 + 64)

 acc1 + = diff1 (i) * 4;

else if (diff1 (i)> 128 + 64 && diff1 (i) <128 + 64 + 32)

 acc1 + = diff1 (i) * 8;

if (diff2 (i) <128)

 acc2 + = diff2 (i);

else if (diff2 (i)> 128 && diff2 (i) <128 + 32)

 acc2 + = diff2 (i) * 2;

else if (diff2 (i)> 128 + 32 && diff2 (i) <128 + 64)

 acc2 + = diff2 (i) * 4;

else if (diff2 (i)> 128 + 64 && diff2 (i) <128 + 64 + 32)

 acc2 + = diff2 (i) * 8 ;, where acc1 is the cumulative weight pixel value of the first pixel difference, diff1 is the first pixel difference, acc2 is the cumulative weight pixel value of the second pixel difference, and diff2 is the zero It's a two-pixel difference.

The image processing apparatus may increase the weight of a portion having a large pixel difference, that is, a weight of a portion having a large movement between fields to facilitate detection of a scene change.

4 is an exemplary diagram illustrating an operation of detecting an editing error of an edited video during an operation of converting the edited video of FIG. 3 into a progressive video.

The value of the field difference varies depending on whether or not the field for which the field difference is obtained belongs to the same frame. If the field for which the field difference is found belongs to the same frame, the field difference is small. The image processing apparatus distinguishes the case where the field at the center of the field difference is f4 or f5, determines whether the field for obtaining the field difference belongs to the same frame, and performs editing error detection differently.

When an image processing apparatus detects an editing error by dividing a case where a field is f4 or f5, the virtual code is expressed as a virtual code.

if (f4)

{

 if (2 * acc2> acc1)

  bbadedit = FALSE;

 else

  bbadedit = TRUE;

}

else if (f5)

{

 if (acc2 <acc1 * 1.5)

  bbadedit = FALSE;

 else

  bbadedit = TRUE;

}, Where acc1 is the cumulative weighted pixel value of the first pixel difference (hereinafter, denoted as the first weight field difference), and acc2 is the cumulative weighted pixel value of the second pixel difference (hereinafter, referred to as the second weighted field difference). ), If bbadedit is FALSE, it performs the operation of detecting the last editing error, and if it is TRUE, it detects the editing error.

The image processing apparatus detects an editing error if the second weighted field difference is smaller than 1/2 * first weighted field difference if the position of the field at the center of the field difference is the last position of the frame, or otherwise detects the last editing error. .

If the position of the field at the center of the field difference is the first position of the frame, the image processing apparatus detects an editing error if the second weight field difference is larger than 3/2 * first weight field difference, or detects the last editing error.

When the operation of detecting the last editing error in the image processing apparatus is expressed in virtual code,

if (acc2> acc1 && acc2 / acc1> = Th)

 bbadedit = TRUE;

else

 bbadedit = FALSE; where acc1 is the cumulative weighted pixel value of the first pixel difference, acc2 is the cumulative weighted pixel value of the second pixel difference, Th is the threshold, and if bbadedit is FALSE, no editing error is detected. If TRUE, editing error is detected.

The image processing apparatus compares the motion difference of acc2 / acc1 with a threshold and detects an editing error if the motion difference is larger than the threshold.

When the image processing apparatus detects an editing error, the image processing apparatus needs to switch from the film mode to the video mode to convert the progressive image using another method without performing the 3: 2 pulldown.

As described in detail above, in the present invention, when the film mode is interrupted when the interlaced video is converted into a progressive video, the present invention detects an image editing part and switches to the video mode without performing a 3: 2 pulldown to display the interlaced video. By converting to a progressive image, the editing screen is accurately detected in the progressive image conversion of the film mode, thereby preventing deterioration of image quality in the film mode.

Claims (4)

Calculating a weighted field difference between front and rear fields based on a predetermined field of an image source; Determining a position of a predetermined field in the frame and performing a case comparison between the weighted field differences of the front and back fields according to the position to detect an editing error; An editing error detection method of the film mode detection method, characterized in that the step of detecting the last editing error by comparing the ratio of the weight field difference of the front and rear fields and the threshold when no editing error is detected. The method of claim 1, wherein the step of calculating the weight field difference is Calculating a difference between pixels of the front and back fields based on one field; Calculating the weighted pixel difference by further comparing the pixel difference values by increasing the weights in proportion to the difference between the pixel difference values; Comprising a step of accumulating the weighted pixel difference for each pixel of the field, the editing error detection method of the film mode detection method. The method of claim 1, wherein the editing error detection step is Detecting an editing error if the second weighted field difference is less than the first weighted field difference if the position of the predetermined field is the last position of the frame; And if the second weighted field difference is greater than the first weighted field difference if the position of the predetermined field is the first position of the frame, the editing error is detected if the second weighted field difference is larger than the first weighted field difference. The method of claim 1, wherein the last editing error detection is Editing error detection method of the film mode detection method, characterized in that for calculating the ratio of the second weighted field difference of the first weighted field difference and comparing the calculated ratio and the threshold, if the calculated ratio is greater than the threshold.

Images