JP2009296042A - Telecine conversion image determination apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a telecine conversion image determination apparatus and method for appropriately determining an encoding system with little processing and performing encoding when still images are inputted in the inverse telecine encoding of telecine images. <P>SOLUTION: When performing the inverse telecine encoding of the telecine images, the magnitude of correlation is determined for pixels between inputted images and the images of the same field inputted immediately before the inputted images. Whether or not the inputted image is the still image is determined using the magnitude of the correlation, and when it is the still image, the inverse telecine encoding is continued. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a telecine-converted image determination apparatus and method for efficiently performing compression encoding by converting a playback moving image composed of frame units with different numbers of images displayed per second.

  A motion picture (cinema video) of a movie film is recognized as a motion picture by the visual afterimage phenomenon of the viewer by continuously displaying images of 24 frames per second (one frame is one image). On the other hand, a television signal based on the NTSC system continuously displays 30 frames of images per second. Furthermore, one frame of the television signal is composed of a combination of two fields. For this reason, a moving image of a television signal is configured by continuously displaying images of 60 fields per second.

  Here, in order to display a cinema video on a television signal receiver of the NTSC system, it is necessary to convert and display a cinema video of 24 frames per second into a television signal of 60 fields per second. This conversion is called “telecine conversion”. That's it.

  As a method for realizing the telecine conversion, there is generally a “2-3 pull-down method (also referred to as 2: 3 pull-down method)”. This is a method in which the first frame of the cinema image is converted into 60 fields by overlapping such that the first frame is 2 fields, the second frame is 3 fields, the third frame is 2 fields, the fourth frame is 3 fields,. There is a 3-2 pull-down method (3: 2 pull-down method) in which the first one of the cinema images corresponds to 3 fields and the second image corresponds to 2 fields, but the 3-2 pull-down method is a 2-3 pull-down method. It is substantially the same processing.

  FIG. 3 is a diagram illustrating telecine conversion. When a cinema video as shown in “Cinema video” in FIG. 3A is input, each cinema video has a top field of cinema video “1” as shown in “Telecine conversion” in FIG. “1t” and bottom field “1b”, and cinema image “2” is converted to top field “2t”, bottom field “2b”, and top field “2t ′” (“′” indicates a duplicate field image) Is done.

  Similarly, the cinema image “3” has a top field “3t” and a bottom field “3b”, and the cinema image “4” has a top field “4t”, a bottom field “4b”, and a bottom field “4b ′” (overlapping field image). ).

  When this is displayed as a television signal, as shown in “Display” in FIG. 3C, “1t” and “1b” are the first frame, “2t” and “2b” are the second frame, “ “2t ′” and “3b” are displayed in the third frame, “3t” and “4b” are displayed in the fourth frame, “4t” and “4b ′” are displayed in the fifth frame, and so on.

  When the telecine conversion is normally performed, it can be considered that patterns P1 to P4 (hereinafter referred to as telecine patterns) as shown in FIG. 4 are continuous. In the telecine pattern shown in FIG. 4, “Bt ′” of P2 and “Db ′” of P4 are overlapping fields. Video (telecine video) is created by the telecine conversion as described above.

  When digital video compression encoding (for example, MPEG-2 or H.264, hereinafter referred to as encoding) is performed on the created telecine video, duplicate field images such as “2t ′” and “4b ′” It is the same as “2t” and “4b”. Therefore, it is possible to improve the encoding efficiency by detecting and deleting the duplicate field image in the telecine video. The process of converting an image of 60 fields / second to an image of 48 fields / second is called inverse telecine conversion.

  Here, an example of the conventional inverse telecine conversion process is shown below.

  As shown in FIG. 3 (d), the overlapping field is detected between top fields (for example, “1t” and “2t”) that are temporally adjacent in a telecine video or between bottom fields (for example, “1b” and “2b”). This is done by examining the correlation of the pixel values of “)”. For example, “2t” and “2t ′” have the same correlation because they are originally the same image.

  In this case, the field “2t ′” determined to have a high correlation is detected, and a determination is made as to whether the detected position of the highly correlated field matches the position of the overlapping field of the telecine pattern (hereinafter, telecine determination). ) Is performed. If the results of the telecine determination match, it is determined that the input video is a telecine video, and the field corresponding to the duplicate field is deleted and encoding (hereinafter referred to as inverse telecine encoding) is performed.

  In such a prior art, when a telecine video still image is input during inverse telecine encoding, it is determined that the correlation between fields is high. Therefore, even if it is not actually a duplicate field, it may be erroneously determined as a duplicate field.

  For example, when a telecine image in which the fifth frame to the tenth frame (section surrounded by a dotted line) as shown in FIG. 5A is a still image 22 is input, the still image section from the fifth frame to the tenth frame. Are substantially the same pixel values between fields, it is determined that the correlation between fields is high.

  As a result of the telecine determination, when the position of the detected highly correlated field and the position of the overlapping field of the telecine pattern in FIG. 4 do not match, it is recognized that the input video is not a telecine video because it is recognized that the input video is out of the telecine pattern. Is done. If it is determined that the input video is not a telecine video, the inverse telecine encoding is stopped, and normal encoding (hereinafter, normal encoding) is performed as shown in FIG. 5B. Since normal encoding is performed without deleting duplicate fields, duplicate fields corresponding to “Bt ′” and “Db ′” in the telecine pattern of FIG. 4 are encoded without being deleted, resulting in a decrease in encoding efficiency. There is a possibility that image disturbance or image degradation may occur.

  After that, when the input of the still image is finished and the moving image is input, it is determined that the input moving image is a telecine video, and in order to resume inverse telecine encoding, it is necessary to match the telecine pattern in the telecine determination. is there. During the telecine determination, since it is not detected that the image is a telecine video, normal encoding is continued. For this reason, duplicate fields (fields 20 and 21 indicated by hatching in FIG. 5B) are not deleted, and coding efficiency decreases.

  On the other hand, a technique for preventing image distortion and image quality degradation even when a telecine video still image is input has been known (see Patent Document 1). This technique generates pull-down information indicating what kind of inverse telecine conversion is performed on an input video after inverse telecine encoding is performed on the input video. The generated pull-down information is corrected, and inverse telecine encoding is performed again based on the corrected pull-down information.

By using such an existing technology, even when a still image is input, the pull-down information is corrected to prevent image distortion and image quality deterioration.
JP 2000-041244 A

  However, in the conventional technique, the pull-down information created from the inverse telecine-encoded video is corrected, and the encoding is performed based on the pull-down information. It was complicated and processing time was long.

  The present invention has been made in view of the above, and an object of the present invention is to appropriately determine an encoding method with a small amount of processing and perform encoding when a still image is input during inverse telecine encoding of a telecine video. An object of the present invention is to provide a telecine conversion image determination apparatus and method.

  In order to achieve the above object, the present invention according to claim 1 is a telecine conversion image determination apparatus for inverse telecine encoding of a telecine video, wherein the input video is newly input and immediately before the input video. Pixel correlation determination means for determining the degree of correlation between pixels constituting both of the input video of the same field and the input video, the input video is a still image based on the determination result by the pixel correlation determination means And inverse telecine conversion means for continuing the inverse telecine encoding in the case of a still image.

  According to a second aspect of the present invention, there is provided a telecine conversion image determination method for inverse telecine coding of a telecine video, the input video newly input by the pixel correlation determination means and the input just before the input video. A step of determining a degree of correlation between pixels constituting both of the input video in the same field and an inverse telecine conversion unit, and the input video is converted into a still image based on a determination result by the pixel correlation determination unit. And continuing the inverse telecine encoding in the case of a still image.

  According to the present invention, when a still image is input during inverse telecine encoding of a telecine video, a telecine conversion image determination apparatus and method for appropriately determining an encoding method with less processing and performing encoding are provided. can do.

  FIG. 1 shows a block diagram according to an embodiment of the present invention.

  FIG. 1 shows a pixel correlation determination 1 in which a video signal is input from the outside, a telecine video determination 3 for receiving a determination output from the pixel correlation determination 1 and performing a telecine video determination, and a determination information storage memory 2. Inverse telecine conversion 4 and encoding process 5 are shown, and the video signal subjected to inverse telecine conversion is output from encoding process 5.

  A video signal input from the outside is input to the pixel correlation determination 1. In the pixel correlation determination 1, the magnitude of the correlation is determined for the pixel between the input video (moving image signal) and the video in the same field input immediately before the input video. The pixel correlation determination result is output to the determination information storage memory 2, and the input video is output to the telecine video determination 3.

  In telecine video determination 3, video determination is performed as to whether the input video is a still image or a telecine video. The telecine video determination 3 requests the determination information storage memory 2 to output pixel correlation determination information when a video is input.

  The determination information storage memory 2 outputs the requested pixel correlation determination information to the telecine video determination 3. The telecine video determination 3 compares the pixel correlation determination information of the input video with the telecine patterns P1 to P4 and confirms whether they match.

  For example, if the pixel correlation determination information matches the telecine pattern, it is determined that the input video is a telecine video. Even if the pixel correlation determination information indicates that the correlation between the pixels is high, if the pixel correlation determination information does not match the overlapping field of the telecine pattern, it is determined that the input video is a still image.

  Thereafter, the telecine video determination unit determines the encoding method based on the information obtained from the pixel correlation determination information. When it is determined that the input video is a telecine video, the inverse telecine conversion is performed by the inverse telecine conversion 4 and then sent to the encoding process 5 where it is encoded and output. In the case of inverse telecine encoding, even when the input video is a still image, as in the case of the telecine video, the inverse telecine conversion is performed by the inverse telecine conversion 4 and then sent to the encoding process 5 for encoding. Is done.

  If the input video is not a still image and does not match the telecine pattern, the input image is not a telecine video, and is sent to the encoding process 5 without being subjected to inverse telecine conversion. .

  Next, referring to FIG. 2, verification of the effect of the embodiment of the present invention will be described.

  It is assumed that inverse telecine encoding is performed on the input telecine video. When an image as shown in FIG. 2 (a) is input and the section 10 surrounded by a dotted line is a still image, the conventional technique differs from the telecine pattern in the still image section, so that the inverse as shown in FIG. 2 (b). Normal encoding is performed without performing telecine encoding. Therefore, in order to encode the still image section of the section 11 surrounded by the dotted line, an encoding process for 5 frames must be performed.

  Further, when a still image is completed and a telecine video is input again, it is necessary to match at least one period of the telecine pattern before the inverse telecine encoding is restarted after the telecine determination. Therefore, in the prior art shown in FIG. 2B, the processing of 5 frames is performed by normal processing after the moving picture of the telecine video is input.

  On the other hand, in the present embodiment, it is determined whether or not the image is a still image, and if it is a still image, the inverse telecine coding method is continued. Therefore, as shown in FIG. Thus, the encoding process for four frames may be performed. Further, when the moving image is resumed, since the inverse telecine encoding is continued for the section where the normal encoding is performed in FIG. 2B, the encoding process for four frames may be performed.

  As described above, according to the present embodiment, since the number of frames to be encoded decreases, the amount of bits allocated to one frame increases, so that the encoding efficiency increases and it is possible to prevent deterioration of image quality due to compression.

  As described above, in this embodiment, when the telecine video is subjected to inverse telecine coding, the magnitude of the correlation is determined for the pixel between the input video and the video of the same field input immediately before the input video. To do. Using the magnitude information of the correlation, it is determined whether the input video is a still image. If it is a still image, inverse telecine coding is continued.

  As a result, it is possible to prevent an increase in the number of frames to be encoded, so that the amount of bits allocated to one frame increases, thereby preventing a decrease in encoding efficiency and a decrease in video quality. At the same time, since the encoding process does not require a two-stage process, the processing time can be shortened as compared with Patent Document 1.

  According to the embodiment of the present invention described above, when a still image is input during inverse telecine encoding of a telecine video, a telecine conversion for appropriately determining an encoding method with a small amount of processing and performing encoding is performed. An image determination apparatus and method can be provided.

The block diagram which concerns on embodiment of this invention is shown. An explanatory view for explaining verification of an effect of an embodiment of the invention is shown. An explanatory view for explaining telecine conversion in the prior art is shown. The explanatory view for explaining the case where telecine conversion was normally performed in the prior art is shown. An explanatory view for explaining a prior art is shown.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Pixel correlation determination 2 ... Determination information storage memory 3 ... Telecine video determination 4 ... Inverse telecine conversion 5 ... Encoding process

Claims (2)

A telecine conversion image determination device for inverse telecine coding of telecine video,
Pixel correlation determination means for determining the degree of correlation for pixels constituting both the newly input video and the input video of the same field input immediately before;
Inverse telecine conversion means for determining whether or not the input video is a still image based on a determination result by the pixel correlation determination means, and continuing the inverse telecine encoding in the case of a still image;
A telecine conversion image determination device comprising: A telecine conversion image determination method for inverse telecine encoding of telecine video,
A step of determining a degree of correlation for pixels constituting both the newly input video and the input video of the same field input immediately before by the pixel correlation determining means;
A step of determining whether the input video is a still image based on a determination result by the pixel correlation determination unit by an inverse telecine conversion unit, and continuing the inverse telecine encoding in the case of a still image;
A telecine conversion image determination method comprising:

Images