JPH10262242A - Compression signal conversion system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the degradation of a resolution by generating an intermediate image between the two frames that always continue and replacing the intermediate image with a skip video image when a skip phenomenon is generated. SOLUTION: A skip scene detection/compensation circuit 8 is inserted between a moving image decoder and a moving image compressor and the frame matching detector 14 of the skip scene detection/compensation circuit 8 compares the 1-frame delay output (1) of a frame delay device 12 with input signals (b) and outputs matching signals in the case that they are the same image. Then, in the case that a skip is generated, even through the matching signals are outputted for the output of the frame matching detector 14, the value of a motion vector detector 13 is kept at a normal value. Based on the result, in a skip scene detector 15, skip scene detection signals are outputted only in the case that the value of the motion vector detector 13 is within a detectable range and the matching signals are outputted. As a result, a switch 17 is switched from (x) and (y) and the output signals of a frame interpolation device 16 are outputted from the terminal of the switch 17z.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is to improve the deterioration of the moving resolution (jerky movement) when converting moving picture signals between different high efficiency coding systems.

[0002]

2. Description of the Related Art In recent years, a high-speed data transmission technique has been remarkably developed in addition to an improvement in a compression processing technique for a moving image signal and an increase in the capacity and speed of a digital recording medium. Further, a digital moving picture editing system integrating them has been put to practical use. FIG. 2 shows an example of a general moving image editing system. Video signals obtained from the television camera 1 and the VTR 2 are compressed by the moving image compressor 3 using the first high-efficiency coding method. The compressed data a output from the moving image compressor 3 is stored in the server 5 via the digital transmission line 4. The compressed data a stored in the server 5 is read out based on a command from the editor 6, decompressed by a moving picture decoder 7 corresponding to the first high-efficiency coding method, and restored to a decoded video signal b. Here, the quality of the decoded video signal b substantially returns to the video signal before compression, but when this video signal is supplied to the recording medium 11 such as the digital broadcasting system 10 or VTR, compression efficiency is more important than editability. The compressed signal compressed by the moving image compressor 9 according to the second high-efficiency coding scheme viewed is used. In general, for the purpose of high-quality editing work such as used by broadcasting stations,
The compressed data a from the server 5 is not used in the digital broadcasting system 10 or the like as it is because the data rate is too high. That is, at present, the compression method suitable for editing and the compression method suitable for transmission do not always match.

[0003] Currently, M-JPEG (Motion-Joint Photographi-
c Experts Group) and MPEG2 (Moving Picture codi
ngExperts Group). Here, both features are summarized as follows regarding “editing work” and “compression efficiency”. 1. M-JPEG: suitable for "editing work", poor "compression efficiency" Because it compresses each frame independently, it is ideal for editing work in frame units. However, a large amount of data is required to maintain a certain level of quality due to poor compression efficiency. 2. MPEG2: “Inappropriate editing work” unsuitable, “compression efficiency” good Compression efficiency is enhanced using inter-frame correlation, and generation of compressed data is also controlled. However, since the inter-frame correlation is used, it is difficult to perform high-quality editing on an arbitrary frame. As a result, the M-JPEG method is most suitable for editing work, and the MPEG-2 method is most suitable when high compression is desired. M-JPEG method suitable for editing work requires a large amount of data (high-speed data) to obtain high-quality real-time images.
is necessary. In particular, in the M-JPEG method, since an entire screen is compressed so as to have a constant image quality, an unpredictable data amount may occur for a complicated image. For this reason, the data transfer capability of the server or the network may be exceeded, and an image skipping phenomenon may occur. This skip phenomenon is that an image of one frame is not reproduced,
Since the immediately preceding image is repeatedly displayed, the dynamic resolution is reduced and the quality is significantly reduced.

[0004]

As described above, in the prior art, when the M-JPEG system with good editability is used,
A skip image may occur depending on the performance of the storage medium or the performance of the transmission path. The occurrence of this phenomenon is unpredictable, and an image that once becomes a skipped image is not improved thereafter. An object of the present invention is to eliminate these drawbacks and reduce the degradation of dynamic resolution in a compressed signal conversion system.

[0005]

In order to achieve the above object, the present invention comprises means for monitoring the continuity of an M-JPEG decoded signal and means for generating an intermediate image between two consecutive frames. When a skip phenomenon occurs, the apparent video resolution is reduced by replacing the skipped video with an intermediate image.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will now be described with reference to FIG.
This will be described in detail with reference to FIG. The same blocks as those in FIG. 2 described above are denoted by the same reference numerals, and description thereof will be omitted. The present invention provides a video decoder 7 of the conventional system shown in FIG.
The maximum effect can be obtained by inserting a skip scene detection / compensation circuit 8 between the image data and the moving image compressor 9. Here, the skip scene detection / compensation circuit 8 of FIG.
This will be described with reference to FIG. In FIG. 3, a decoded video signal b of the first high-efficiency coding system (M-JPEG) is applied to a frame delay unit 12, a motion vector detector 13, and a frame coincidence detector 14. The two-frame delay output c of the frame delay unit 12 is connected to the x terminal of the switch 17, the motion vector detector 13, and the frame interpolator 16. The output of the motion vector detector 13 is applied to a frame interpolator 16 and a skip scene detector 15. Further, the output of the frame interpolator 16 is
Of the skip scene detector 15
Is controlled by the output. Note that the frame coincidence detector 14 compares the one-frame delayed output of the frame delay unit 12 with the input signal b, and outputs a coincidence signal when the images are the same. As a detection method, a difference between the two image signals is obtained, and when the difference is smaller than a predetermined value, a method of outputting a coincidence signal. Alternatively, the amount of a motion vector between the two is detected and the value is small. Various detection methods, such as outputting a coincidence signal in such a case, can be applied.

Next, the operation of these circuits will be described. The motion vector detector 13 is connected to the input decoded video signal b of the frame delay unit 12 and the output c delayed by two frames. Here, horizontal / vertical vectors are detected by a motion vector detection method widely used in an encoding method such as MPEG. As an example of a specific image, FIG.
An image b of the pendulum in the frame is shown as an image of a pendulum in the (n-1) th frame and an image c of the pendulum in the (n-2) th frame. Note that the flow of time has been n-2, n-1,
It goes without saying that the order is n. here,
In the case of “no skip”, the frame match detector 1
As the output of No. 4, a non-coincidence signal is output, and the value of the motion vector detector 13 is also a normal value. On the other hand, in the case of "there is a skip", the output of the frame match detector 14 outputs a match signal, but the value of the motion vector detector 13 remains a normal value. Based on this result, the skip scene detector 15 outputs a skip scene detection signal only in a range where the value of the motion vector detector 13 can be detected and only when a coincidence signal is output.

As a result, the switch 17 is switched from x to y, and the output signal of the frame interpolator 16 is output from the z terminal of the switch 17. Although no example is given here, n-2 and n
Obviously, since there is no correlation between the two frames of the frame, it is impossible to detect a motion vector. Therefore, when the motion vector cannot be detected, it is a change part of the scene, and when the motion vector shows a gradual value,
It can be determined as a continuous scene. The output d of the frame interpolator 16 connected to the y side of the switch 17 is
2 is a signal generated by using the value of the motion vector detector 13 as a reference and generating an image d of n-1 frames.
Note that the frame interpolator 16 may average the input b of the frame delay unit 12 and the output c delayed by two frames as a simple method without using the motion vector as described above. Blurring occurs with the formation. Furthermore, although the description is omitted here, when the input of the system is a film video or the like, a process called 2-3 pulldown is performed, so that a periodic frame match is detected. In this case, it is necessary to judge the film material and stop the interpolation.

[0009]

According to the present invention, high-precision interpolation can be performed in a portion where one frame has been omitted, and it is possible to efficiently perform frame skipping caused by the compression process of M-JPEG and insufficient performance of a transmission circuit. A system that can perform compensation and improve the quality of moving images can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the overall configuration of the present invention.

FIG. 2 is a block diagram showing an example of a conventional system.

FIG. 3 is a block diagram showing an example of a main part of the present invention, a skip scene detection / compensation circuit.

FIG. 4 is a diagram for explaining a motion vector detection method.

[Explanation of symbols]

1: TV camera, 2: VTR, 3: moving image compressor,
4: transmission path, 5: server, 6: editor, 7: video decoder, 8: skip scene detection / compensation circuit, 9: video compressor, 12: frame delayer, 13: motion vector detector, 14: frame match detector, 15: skip scene detector, 16: frame interpolator, 17: switch.

Claims (1)

[Claims] 1. A compressed signal conversion system for decompressing an image compressed by a first high-efficiency encoding system and compressing the image again in accordance with a second high-efficiency encoding system. Means for detecting a skip scene from a decompression signal of an image according to a method, and interpolation means for generating an image corresponding to a skip scene from scenes before and after the detected skip scene, and when a skip scene is detected,
A compressed signal conversion system characterized in that a signal having a temporally continuous dynamic resolution is obtained by inserting an interpolated image generated in the portion.