JP2007288483A - Image converting apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the image quality of a display video when outputting an input video signal and an interpolated video signal by detecting motion vector information from the input video signal and generating the interpolated video signal by using the motion vector information. <P>SOLUTION: A video selecting part 104 inputs the input video signal, motion vector information detected by a motion vector detecting part 101 and an interpolation phase generated by a field interpolation phase generating part 102, detects a difference between a phase of the input video signal and the interpolation phase, and selects and outputs the input video signal in place of the interpolated video signal when the phase difference is smaller than a predetermined level. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image conversion apparatus for generating an output signal composed of an input video signal and a new video signal generated using motion vector information detected from the input video signal, and in particular, a telecine apparatus performs 3-2 conversion. The present invention relates to an image conversion apparatus capable of obtaining a high-quality television signal by performing video interpolation for motion vector detection and motion correction on a television signal obtained by 2-2 conversion. is there.

  In recent years, liquid crystal panels have been widely used as display devices for TV. However, due to the poor response performance, blurring when displaying moving images has been a problem. On the other hand, the response characteristics of the panel itself have been remarkably improved, and the moving image blur as before has disappeared recently. However, due to the display method of the liquid crystal itself that keeps displaying the same picture for a certain period of time, it is known that even if the response speed of the liquid crystal panel itself is improved, the motion blur will not disappear in principle. It has been. In contrast, efforts are being made to improve motion blur by shortening the retention period of the video signal to be displayed.

  In order to enable such display, it is necessary to generate and interpolate a video signal between the input video signal and the video signal, and convert the frame rate of the output video signal to, for example, 1.5 times or 2 times. . This frame rate conversion process detects the motion of the video on the entire screen or for each individual area by detecting the correlation between a plurality of video signals input in series, and this motion information, so-called motion A general method is to generate an interpolated video signal that interpolates the motion of the preceding and succeeding video signals between the input video signal and the video signal using a vector.

  FIG. 2 shows the relationship between an input signal and an output signal when a video signal of 60 fields per second is converted to 90 fields per second by frame rate conversion processing.

  On the other hand, when a movie film is used as a material for transmitting or recording / playing moving image data, a telecine device is used to convert a movie film source of 24 frames per second into an interlaced video signal of 60 fields per second using a 3-2 pull-down method. Scan conversion may be performed.

The 3-2 pull-down method is one of methods used when converting (pull-down) video recorded at 24 frames per second such as movie film into a video signal of 60 fields per second. Each frame of the film is converted into 2 fields for odd frames and 3 fields for even frames, or vice versa, thereby converting 2 frames to 5 fields, that is, 24 frames to 60 fields. FIG. 3 (a) shows a movie film source of 24 frames per second, and FIG. 3 (b) shows the result converted into 60 fields by the 3-2 pull-down method. When converted by the above method, there is an image that should be reproduced at the same time interval of 1/24 second, one is 3/60 seconds, and the next image is 2/60 seconds. Therefore, there is a problem that an unnatural movement occurs as a moving image. Therefore, a method has been proposed in which a motion vector of a video signal is detected, and a frame is synthesized and inserted at a correct position corresponding to a time difference to make the motion of the displayed video natural.
Japanese Patent No. 3495485

  However, the image quality of the video generated by the image conversion apparatus as described above greatly depends on the accuracy of the detected motion vector. Here, there is a region where a motion vector cannot be detected in principle, such as a portion where a moving object and the background intersect, and there is a case where a correct motion vector cannot always be detected due to restrictions on the processing capability of the vector detection unit. Therefore, it is necessary to assume that an image combined with a motion vector is not necessarily ideal.

  In particular, when the original data is converted from a movie film source of 24 frames per second into a video signal of 60 frames per second, as is apparent from FIG. 3C, from the originally input video in the output signal. It can be seen that the ratio between the directly generated frame and the frame synthesized using the motion vector is 1: 4. That is, the synthesized frame is displayed for a period of 80% within a certain time, and this is one of the causes of poor image quality of the display image.

  Therefore, an object of the present invention is to provide an image conversion apparatus that solves the above-described problems.

  In order to achieve the above object, the invention according to claim 1 is directed to a motion vector detecting means for detecting motion vector information from an input video signal, and a temporal generation time for generating an interpolated video between the input video signals. In an image conversion apparatus comprising: an interpolation phase generation unit that generates phase information; an interpolation video generation unit that generates an interpolation video signal from the video signal, the motion vector information, and the interpolation phase information; The input video signal and the interpolated video signal are selectively switched and output based on the interpolation phase information.

  The invention according to claim 2 relates to the invention according to claim 1, wherein a temporal phase difference between the input video signal and the interpolated video signal is determined from the interpolated phase information, and The input video signal is output in place of all or a part of the interpolated video signals having a phase difference smaller than a predetermined level.

Further, the invention according to claim 3 relates to the invention according to claim 1 or 2, and whether or not the input video signal is input and the input video signal is an image obtained by 3-2 conversion or 2-2 conversion. Further, the interpolation phase generation unit outputs interpolation phase information based on a detection result of the film detection unit.

  As described above, with the image conversion apparatus of the present invention, the image quality when generating the interpolated video signal using the motion vector information, particularly the image obtained by converting the input video signal to 3-2 conversion and 2-2 conversion It is possible to improve the image quality when generating and outputting the interpolated video signal.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Embodiment 1)
First, an image conversion apparatus according to a first embodiment of the invention will be described with reference to FIGS. 1 and 3. FIG. 1 is a block diagram showing a configuration of an image conversion apparatus according to the first embodiment of the present invention. As shown in FIG. 1, a motion vector detection unit 101, a field interpolation phase generation unit 102, a field interpolation are shown. The video generation unit 103 and the video selection unit 104 are configured.

  The field interpolation phase generation unit 102 first determines whether or not the input video signal is a 3-2 converted or 2-2 converted image based on the difference between successive field image data. For example, in the case of the 3-2 converted video signal shown in FIG. 3B, the timing of the current input signal from b0 to b4 is specified. This means that even if the video itself is a moving image, b0, b1, and b2, and b3 and b4 in FIG. 3B are displayed at the same position on the screen, focusing on the fact that the original data is the same video. This is performed by detecting a period in which no difference occurs between the video signals to be processed. Of course, when the video signal shown in FIG. 3B is in the interlace format, for example, a difference occurs between b0 and b1, and it is necessary to use a determination unit that allows a certain amount of difference. Since this determination means is not related to the essence of the present invention, the details are omitted. Here, it is assumed that the detection of the sequence 3-2 is correctly performed. When the input video signal is a signal that has undergone 3-2 conversion, the field interpolation phase generation unit 102 further determines at which timing in the sequence 3-2 or 2-2 as shown in FIG. It detects whether there is any and calculates the interpolation phase.

  For example, if it is found that the current input is b1 in FIG. 3B, it is generated as c1 in FIG. 3C at a position of 20% as a temporal phase with respect to b0 and b5. Will be.

  As a result, as shown in FIG. 3C, it is possible to accurately reproduce the motion of the object that is moving at a constant speed. In order to generate such an interpolated image, motion vector information detected by the motion vector detection unit 101 is required. The motion vector detection unit 101 generally has a memory capable of storing video signals of a plurality of fields, and detects a motion vector for each pixel or a plurality of pixels based on the difference between the video data of the plurality of fields. For example, motion vectors for generating the images c1 and c2 in FIG. 3C are detected from a0 and a2.5 in FIG. If the entire screen is moving uniformly and 5 pixels are detected in the right direction of the screen and 5 pixels in the upward direction of the screen, the above pixel on c0 is moved 2 pixels on the right and 2 pixels on the right. Thus, the picture of c1 is generated. Similarly, c3 and c4 are generated using a motion vector detected between a2.5 and a5. The motion vector detection method is a known technique and will not be described in detail here.

  Here, for example, an image generated as c3 is obtained by moving a2.5 one pixel to the right and up. If the detected vector is ideal, c3 is also an ideal image that smoothly connects with the preceding and following images, but as described above, the detected vector includes errors and errors, An image generated using the image may be different from an ideal one. Therefore, as shown in FIG. 3D, by using b3 instead of c3, an image independent of the accuracy of the vector is generated.

  Here, when the phase difference between a0 and a5 in FIG. 3 (a) is 10, the temporal phase difference between c3 in FIG. 3 (c) and d3 in FIG. Become. Thus, when the phase difference between the interpolated image and the input image is small, the influence on the smoothness of the movement is small, and the visual discomfort is also small. At the same time, the phase difference between c2 and b3 in FIG. However, if d2 is replaced with b3, the smoothness of movement is significantly impaired, which is not preferable.

  In this way, when the video is moving at a uniform speed due to the replacement of the interpolated image and the input image, the smoothness of movement between the previous and next videos is impaired, but compared with FIG. 3B. Thus, the improvement in the overall smoothness of the movement is obvious, and in FIG. 3C, the image generated depending on the vector is continued for 4 fields, whereas the maximum of 2 fields is continued. Thus, it is possible to reduce image quality degradation when the vector accuracy is poor.

  In FIG. 3, the case where b3 is used instead of c3 as d3 has been described, but b2 may be used instead of c2 as d2.

  FIG. 3 shows a case where the original 24 frames of video are converted into 60 fields. However, when the original 25 frames of video are converted into 50 fields, the 60 fields of video are 90 fields and 120. The same applies to the case of conversion to a field.

(Embodiment 2)
In the first embodiment, the input signal and the interpolated image are fixedly selected depending on only the field interpolation phase. However, in the second embodiment, the output switching is changed to the field interpolation phase. In addition, it is performed depending on the reliability of the vector.

  FIG. 4 shows the configuration of an image conversion apparatus according to the second embodiment of the present invention. The motion vector detection unit 401 newly outputs vector reliability information, and the video selection unit 404 newly inputs this vector accuracy information.

  Motion vector detection for pixels on the screen or subdivided individual areas is performed by detecting a highly correlated part between the previous and next video images. In general, attention is paid to a large difference in signal difference. In some cases, a highly correlated pixel or region cannot always be detected due to the limit of the detection performance of the motion vector detection unit 401 or the motion of the preceding and following images. In such a case, the difference between the images becomes large. The difference information is accumulated in the area and passed to the video selection unit 404 as information indicating the reliability of the detected motion vector.

  The video selection unit 404 refers to the vector reliability information at the same time as the interpolation phase from the field interpolation phase generation unit 102, and determines that the vector reliability is low. Instead, the input video signal is output.

  The motion vector detection unit 401 may output the vector reliability signal as one information for the entire screen, and the video selection unit 404 may switch the entire screen between an input image and an interpolated image, or the motion vector The detection unit 401 may output the vector reliability signal as information for each pixel or region unit, and the video selection unit 404 may switch between the input image and the interpolation image for each pixel or region.

  The present invention relates to an image conversion apparatus that can detect a motion vector and perform video interpolation for motion correction to obtain a high-quality television signal.

1 is a block diagram showing the configuration of an image conversion apparatus according to a first embodiment of the present invention. The figure explaining operation | movement of the image conversion apparatus concerning the 1st Embodiment of this invention. The figure explaining operation | movement of the image conversion apparatus concerning the 1st Embodiment of this invention. The block diagram which shows the structure of the image conversion apparatus concerning the 2nd Embodiment of this invention.

Explanation of symbols

101, 401 Motion vector detection unit 102 Field interpolation phase generation unit 103 Field interpolation video generation unit 104, 404 Video selection unit

Claims (3)

Motion vector detection means for detecting motion vector information from the input video signal;
Interpolation phase generation means for generating temporal phase information for generating an interpolated video between input video signals;
In an image conversion apparatus comprising: the video signal; the motion vector information; and an interpolation video generation means for generating an interpolation video signal from the interpolation phase information.
An image conversion apparatus characterized by selectively switching between the input video signal and the interpolated video signal based on the interpolation phase information. A temporal phase difference between the input video signal and the interpolated video signal is determined from the interpolated phase information, and all the interpolated video signals whose phase difference is smaller than a predetermined level, or some of The image conversion apparatus according to claim 1, wherein the input video signal is output instead of the interpolated video signal. The image processing apparatus further includes film detection means for inputting the input video signal and detecting whether the input video signal is a 3-2 converted or 2-2 converted image, and the interpolation phase generating unit is configured to detect the film detection. 3. The image conversion apparatus according to claim 1, wherein interpolation phase information is output based on a detection result of the means.

Images