JPH057356A - Image recording/reproducing device - Google Patents

Abstract

PURPOSE:To convert an encoded moving image into a still image data so as to edit/reproduce and record in common medium. CONSTITUTION:This device is provided with a moving image input means 20 for inputting a moving image signal, a data compressing means 2 for data- compressing an inputted moving image signal by a high efficient encoding, a recording means 1 for recording encoded moving image and still image signals, an expanding means 3 for decoding encoded moving image and still image signals and a convering means 4 for converting an encoded moving image signal into a still picture signal. To encode a core frame of a moving image signal, the same encoding method as that for a still image is used.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image recording / reproducing apparatus for compressing and recording still images and moving images, and reproducing and displaying them.

[0002]

2. Description of the Related Art On a personal computer, so-called multimedia has come to be used which can freely handle various kinds of information such as image signals and voice signals in addition to text data. In such an apparatus, it is desirable that both still images and moving images can be handled as images.
Further, when an image is stored as digital data, the amount of data becomes enormous. Therefore, in order to record a large amount of image information in a limited storage capacity range, it is necessary to perform some compression on the image signal. As a highly efficient compression method of image data, an encoding method combining orthogonal transformation and variable length encoding is widely known. For example, the outline of a method (JPEG method) being studied as an international standard of a still image compression method is as follows.
First, image data is divided into blocks of a predetermined size (blocks of 8 × 8 pixels), and each divided block is subjected to a two-dimensional DCT (discrete cosine transform) as orthogonal transformation.
Do. Next, linear quantization is performed according to each frequency component, and Huffman coding is performed on this quantized value as variable length coding. At this time, with respect to the DC component, the difference value from the DC component of the neighboring block is Huffman coded.
The AC component is scanned from a low frequency component to a high frequency component called a zigzag scan, and Huffman coding is performed on a set of a continuous number of invalid (value 0) components and a value of the following valid component. ..

As the moving picture compression method, the MPEG method is under study as an international standard. That is, in this method,
Intra-frame compression (compression as a still image) and inter-frame compression (compression using prediction between consecutive frames) are combined. In the intraframe compression, basically, similar to the JPEG method described above, the orthogonal transformation and the variable length coding are combined to perform the coding. A frame on which such a still image is compressed is called a core frame, and is arranged at an appropriate interval between consecutive frames. Such a frame is called an I picture.

On the other hand, in inter-frame compression, a predicted value is calculated from a nearby reference frame, and the difference between the frame to be coded and the predicted value is coded by combining orthogonal transformation and variable length coding. In order to obtain a predicted value from nearby frames, it is necessary to detect the motion of the screen and make a prediction. That is, the positional relationship between the reference frame and the coded frame is determined so as to be most similar for each block having a predetermined size, and the positional deviation is used as a motion vector. The motion vector is also encoded as part of the code.

There is a method of using not only past frames but future frames as reference frames. That is, three types of prediction methods are conceivable as motion-compensated frame predictions: forward prediction by past frames, backward prediction by future frames, and average value prediction of both.
In the MPEG system, a frame (P
Picture), forward prediction, backward prediction,
A frame (B picture) is used that selects and uses the optimum prediction from the three prediction values of both average value predictions.

[0006]

As described above, in the image recording / reproducing apparatus used for multimedia,
It is desired that image data can be compressed with high efficiency.
The above-mentioned JPEG is used as a compression method satisfying such requirements.
There are methods and MPEG methods. However, although the coding method for the still image and the coding method for the moving image have the same basic parts, the conversion between the two is not considered. Therefore, for example, an image encoded as a moving image can be reproduced only as a moving image, and it is impossible to extract a frame in the moving image as a still image and edit / record / reproduce it. This impairs the advantage of multimedia, which is originally capable of handling various kinds of information freely.

Therefore, the object of the present invention is to eliminate the above drawbacks and to convert encoded moving images into still image data so that they can be edited / played back and recorded on a common medium. An object of the present invention is to provide an image recording / reproducing device capable of performing the same.

[0008]

In order to achieve the above object, the image recording / reproducing apparatus of the present invention uses a moving image input means for inputting a moving image signal and a high efficiency encoding of the input moving image signal. Data compression means for data compression, recording means for recording the encoded moving image signal and the encoded still image signal, and decoding the encoded moving image signal and the encoded still image signal Decompressing means and conversion means for converting the encoded moving image signal into a still image signal,
When the core frame of the moving image signal is encoded, it is encoded by the same method as the encoding of the still image.

[0009]

That is, in the present invention, when the core frame of a moving picture signal is coded, it is coded by the same method as the coding of a still picture, and still picture data is obtained by format conversion. Is.

[0010]

EXAMPLE An example of the present invention using the above principle will be described. FIG. 1 shows a block diagram of a multimedia system including an image recording / reproducing apparatus according to the present invention. Reference numeral 1 denotes a memory in which image data is recorded, which is composed of a magnetic disk recording device. Reference numeral 2 is an encoding circuit for compressing image data, 3 is a decoding circuit for expanding compressed moving image data, 4 is a format conversion circuit, and 5 is a monitor for displaying an image.

As shown in FIG. 2, the encoding circuit 2 has an input terminal 20 for inputting a moving image signal, an A / D converter 21 for converting an analog signal into a digital signal, and a frame for storing a digitized signal. Memory 22, code amount prediction circuit 2
3, subtraction circuit 24, DCT quantization circuit 25, Huffman coding circuit 26, inter-frame prediction circuit 27, adaptive selection circuit 28, inverse quantization IDCT circuit 29, buffer memory 30
With. The code amount prediction circuit 23 calculates the generated code amount of the frame to be encoded, and sets the quantization width based on this. Orthogonal transform (DCT) quantization circuit 25
Performs DCT on the block-processed image data output from the frame memory 22 as an orthogonal transform for each block, and further uses a quantization width preset for each frequency component for each frequency component. Performs linear quantization. The Huffman coding circuit 26 is for Huffman coding the quantized transform coefficient. The inter-frame prediction circuit 27 determines a motion vector between the reference frame and the encoded frame. The adaptive selection circuit 28 selects a reference frame according to the type of frame to be encoded (I, P, B picture). The inverse quantization IDCT circuit 29 converts the quantized DCT coefficient into a representative value, and further performs inverse DCT conversion to restore the image data. The buffer memory 30 stores Huffman-encoded data.

The operation of the image recording / reproducing apparatus configured as described above will be described. For example, a video signal of a moving image is input to the input terminal 20 from a television camera, converted into a digital signal by the A / D converter 21, and then the frame memory 2
Stored in 2. The stored image data is block-processed into a size of 8 × 8 from the frame memory 22 and input to the subtraction circuit 24. For the core frame, 0 is subtracted (that is, the original signal remains unchanged) and is output to the DCT quantization circuit 25. In the DCT quantization circuit 25, after performing DCT for each block, linear quantization is performed using the quantization width for each frequency component. Here, as the quantization width, a prediction value by the code amount prediction circuit 23 is used according to the data amount information accumulated in the buffer memory 30 so that excess or deficiency of data does not occur.
The code amount prediction circuit 23 obtains an activity (parameter indicating the amount of high frequency component) for the data from the frame memory 22. Since the activity has a correlation with the code amount, the optimum quantization width for encoding the data without excess or deficiency is set and used from this value.

The quantized transform coefficient is Huffman coded by the Huffman coding circuit 26. At this time, regarding the DC component DC, the difference value from the DC component of the neighboring block is grouped according to its bit length, and the Huffman coding indicating the group and the difference value are combined to form coded data. The AC component AC scans from a low frequency component to a high frequency component called a zigzag scan, and the number of consecutive zero-valued components (number of runs of zero) and the group number of the non-zero component values that follow. Huffman coding is performed on the set and the obtained codeword and additional bits are combined to form coded data.

The above processing is sequentially performed for each block until the processing of blocks for one frame is completed. Here, when the core frame is compressed, all blocks for one frame are encoded while maintaining the same quantization width.

Next, when the encoded frame is a P picture, the inter-frame prediction circuit 2 from the frame memory 22.
The encoded frame data is input to 7, and a motion vector between the frame and the reference frame is obtained. The preceding I picture is used as the reference frame. Ie DCT
After that, the linearly quantized data is converted into a representative value by the inverse quantization IDCT circuit 29, and the image data restored by the inverse DCT conversion is input from the adaptive selection circuit 28 to the interframe prediction circuit 27. Interframe prediction circuit 2
In 7, a positional relationship that maximizes the correlation value is obtained for each block of a predetermined size between a P picture that is a coded frame and an I picture that is a reference frame, and the positional deviation is obtained as a motion vector, Encode the vector. After that, the inter-frame prediction circuit 27 performs motion compensation on the I picture that is the reference frame with the obtained motion vector, and outputs it to the subtraction circuit 24. In the subtraction circuit 24, a P picture which is an encoding frame and an I which is a reference frame
The picture and the picture are subtracted and output to the DCT quantization circuit 25.

In the DCT quantization circuit 25, after performing DCT for each block, linear quantization is performed using the quantization width for each frequency component. The quantized transform coefficient is Huffman coded by the Huffman coding circuit 26 without distinction between DC and AC.

When the coded frame is a B picture, the frame memory 2 is used as in the case of the P picture.
Image data is input to the inter-frame prediction circuit 27 from 2 and the motion vector between the reference frame and the reference frame is obtained.
As a reference frame, I picture or P before and after it
Picture is used. The restored image data of the corresponding reference frame is input from the adaptive selection circuit 28 to the inter-frame prediction circuit 27. In the inter-frame prediction circuit 27, from the B picture that is the encoded frame and the I picture or P picture that is the reference frame, the past frame,
Three types of motion vectors of the future frame and the average value of both are calculated, and the optimum prediction is selected according to the evaluation value. The selected reference frame is motion-compensated by the obtained motion vector of the optimum prediction, and is output to the subtraction circuit 24.
The subtraction circuit 24 subtracts the B picture, which is the encoded frame, from the reference frame, and outputs the subtracted signal to the DCT quantization circuit 25. In the DCT quantization circuit 25, after performing DCT for each block, linear quantization is performed using the quantization width for each frequency component. The quantized transform coefficient is Huffman coded by the Huffman coding circuit 26 as in the case of the core frame.

The above processing is sequentially performed to code the input moving image signal and add a data header defined by the MPEG system to the coded data.
That is, a file in the MPEG system has a structure as shown in FIG. 3A, but here, a header relating to an image format including the number of pixels and a quantization matrix is added to the video sequence. The editing information is added to the group of pictures, the reproduction order and the coding type are added to the pictures, the quantization scale factor is added to the slices, and the position and the coding type headers are added to the macroblocks. The coded data with the header added is transferred from the buffer memory 30 to the memory 1 (FIG. 2) and recorded.

Next, the operation when a frame in the compressed moving image data is used as a still image will be described. The compressed moving image data recorded in the memory 1 is sent to the decoding circuit 3 and reproduced by the next operation. That is, in the circuit configuration of the encoding circuit 3 shown in FIG.
The data temporarily stored in the buffer memory 31 in the decoding circuit is sent to the Huffman decoding circuit 32, and the DC component and the AC component are decoded. Decoded data is inverse quantization ID
The quantized DCT coefficient is converted into a representative value in the CT circuit 33, and is further subjected to inverse DCT conversion. In the case of the core frame, the restored data is output as it is to the adaptive selection circuit 34. In the case of a P picture or a B picture, the addition circuit 35 adds the reference frame data. Here, as a reference frame, an I-picture or a P-picture that has already been decoded is output from the adaptive selection circuit 34, and is motion-compensated in the motion compensation circuit 36 to be used.

The image decoded as described above is reproduced on the monitor 5, but if there is an image to be extracted as a still image, the core frame corresponding to the image is selected. The encoded data of the core frame is output from the buffer memory 31 to the format conversion circuit 4. The format conversion circuit 4 makes Huffman-encoded core frame data compatible with the JPEG standard.
That is, as shown in FIG. 3B, the JPEG file defines an SOI code indicating the beginning of image data, a frame header including coding type and pixel number information, a DQT code defining a quantization matrix, and a Huffman code table. DHT code, scan header showing correspondence of color components,
Since the EOI code indicating the end of the image data is required, these headers are created and output to the memory 1.

On the other hand, for the encoded data, the quantized transform coefficient is represented by Huffman code. The quantized transform coefficient is Huffman coded by the Huffman coding circuit 26. At this time, regarding the DC component DC, the difference value from the DC component of the neighboring block is grouped according to its bit length, and the Huffman coding indicating the group and the difference value are combined to form coded data. The AC component AC scans from a low frequency component to a high frequency component called a zigzag scan, and the number of consecutive zero-valued components (number of runs of zero) and the group number of the non-zero component values that follow. Huffman coding is performed on the set and the obtained codeword and additional bits are combined to form coded data. In this way, the JPEG data structure is MPEG
Since it is the same expression as the core frame of, it can be easily converted.

The still image compressed data thus recorded in the memory 1 complies with the standard for still image compression, and can be edited and reproduced in exactly the same manner as the data compressed from a normal still image. .. In the above description, the code amount prediction circuit 23 uses the activity to set the optimum quantization width, but this is not limited to the activity. For example, the actual code amount is calculated and the optimum quantization width is obtained. The width may be set.

Next, a second embodiment of the present invention will be described. FIG. 5 shows a block diagram of the image recording / reproducing apparatus of the second embodiment. In this example, in addition to the memory 1 including a magnetic disk recording device, a memory card 40 having a built-in semiconductor memory is added as a memory for recording image data. The memory card 40 is provided in the main body of the image recording / reproducing device. It is removable from the socket. The memory card 40 is used as a recording medium of an electronic camera that captures a still image. Reference numeral 2 is an encoding circuit for compressing image data, 3 is a decoding circuit for expanding compressed moving image data, 4 is a format conversion circuit, and 5 is a monitor for displaying an image. Reference numeral 6 is a decoding circuit for expanding the still image data. Since the internal configurations of the encoding circuit and the decoding circuit are the same as those of the first embodiment, their description will be omitted.

The operation of the image recording / reproducing apparatus of FIG. 5 will be described with reference to FIG. The video signal of the moving image is input and A
The signal is converted into a digital signal by the / D converter 21 and stored in the frame memory 22. The stored image data is block-processed from the frame memory 22 and input to the subtraction circuit 24. The core frame is output to the DCT quantization circuit 25 as it is as an original signal. In the DCT quantization circuit 25, after performing DCT for each block, linear quantization is performed using the quantization width for each frequency component. Here, as the quantization width, the buffer memory 3
According to the data amount information stored in 0, the quantization width is set and used so that excess or deficiency of data does not occur.
The quantized transform coefficient is Huffman coded by the Huffman coding circuit 26. The above process is sequentially performed for each block until the process for one frame of blocks is completed. Here, when compressing the core frame, all blocks for one frame are encoded with the same quantization width.

Next, when the coded frame is a P picture, the inter-frame prediction circuit 2 from the frame memory 22.
The encoded frame data is input to 7, and a motion vector between the frame and the reference frame is obtained. The preceding I picture is used as the reference frame. The inter-frame prediction circuit 27 performs motion compensation on the I picture that is the reference frame with the obtained motion vector, and outputs it to the subtraction circuit 24. The subtraction circuit 24 subtracts the P picture, which is the encoded frame, from the I picture, which is the reference frame, and D
It is output to the CT quantization circuit 25. DCT quantization circuit 2
In 5, after DCT is performed for each block, linear quantization is performed using the quantization width for each frequency component. The quantized transform coefficient is sent to the Huffman coding circuit 26.
Is Huffman-encoded as in the core frame.

When the encoded frame is a B picture, the frame memory 2 is used as in the case of the P picture.
Image data is input to the inter-frame prediction circuit 27 from 2 and the motion vector between the reference frame and the reference frame is obtained.
As a reference frame, I picture or P before and after it
Picture is used. The selected reference frame is motion compensated by the motion vector of the obtained optimum prediction,
It is output to the subtraction circuit 24. The subtraction circuit 24 subtracts the B frame, which is the encoded frame, from the reference frame,
It is output to the DCT quantization circuit 25. In the DCT quantization circuit 25, after performing DCT for each block, linear quantization is performed using the quantization width for each frequency component. The quantized transform coefficient is sent to the Huffman coding circuit 26.
Is Huffman-encoded as in the core frame.

The above-described processing is sequentially performed, and the input moving image signal is encoded. The encoded data is MPE
A data header specified by the G method is added and recorded from the buffer memory 30 to the memory 1 (FIG. 5). Further, here, the compressed moving image data may be transferred to and recorded in the memory card 40. The memory card 40 has a recording capacity of several megabytes and can record a moving image for several seconds. In recording the encoded moving image data in the memory card 40, if recording is performed for each group of pictures, which is a unit of data structure, recording, editing, reproduction, etc. can be performed efficiently.

On the other hand, the operation when a frame in the compressed moving image data is used as a still image will be described. The compressed moving image data recorded in the memory 1 is temporarily stored in the buffer memory in the decoding circuit 3. The image decoded by the decoding circuit 3 is reproduced on the monitor 5, and the core frame corresponding to the image extracted as a still image is selected. The encoded data of the core frame is output from the buffer memory to the format conversion circuit 4. The format conversion circuit 4 makes Huffman-encoded core frame data compatible with the JPEG standard, creates a data header including quantization width information, and outputs the data header to the buffer memory 30. The obtained still image code data is transferred and recorded in a predetermined memory card.

In this embodiment, a memory card, which is generally used as a recording medium for electronic cameras, is used as the recording medium. Therefore, by inserting the memory card and decoding it by the decoding circuit 6, the image taken by the electronic camera can be easily reproduced on the monitor 5. It is also possible to convert frames in a moving image into still images and record them on a memory card for still images. Further, the compressed moving image can be recorded in a memory card, and the characteristics of multimedia can be further utilized.

As is clear from the above, in the present invention, when the core frame of the moving image signal is encoded by the same method as the encoding of the still image, the still image data is converted by the format conversion. obtain. For example, when a moving image is compressed by the MPEG method, keeping the quantization width constant within a frame is not considered in the normal MPEG method, but in the present invention, encoding is performed without changing the quantization width when compressing a core frame. As a result, still image data compatible with the JPEG system can be obtained.

[0031]

As described above, in the image recording / reproducing apparatus of the present invention, a frame in a moving image can be recorded, edited and reproduced exactly like data compressed from a normal still image. You can According to the method of the present invention, since the moving image is converted to the still image in the encoded data, the image quality is deteriorated due to repeated compression as in the case where the moving image is once decoded and then compressed again as the still image. Does not happen.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an image recording / reproducing apparatus according to the present invention.

FIG. 2 is a configuration diagram of an encoding circuit of the image recording / reproducing apparatus of the present invention.

FIG. 3A is a diagram showing a data file structure in the MPEG system, and FIG. 3B is a diagram showing a data file structure in the JPEG system.

FIG. 4 is a block diagram of a decoding circuit of the image recording / reproducing apparatus of the present invention.

FIG. 5 is a configuration diagram of an image recording / reproducing apparatus according to a second embodiment of the present invention.

[Explanation of symbols]

1 ... Memory, 2 ... Encoding circuit, 3 ... Decoding, 4 ... For
Matt conversion circuit, 5 ... Monitor, 6 ... Decoding circuit.

Claims (1)

Claims: 1. A moving image input means for inputting a moving image signal,
Data compression means for compressing the input moving image signal by high-efficiency encoding, recording means for recording the encoded moving image signal and the encoded still image signal, and the encoded moving image signal And a decompressing unit for decoding the encoded still image signal, and a conversion unit for converting the encoded moving image signal into a still image signal, and when the core frame of the moving image signal is encoded, An image recording / reproducing apparatus characterized by encoding by the same method as image encoding.