JPH06113279A - Picture compressor - Google Patents

Abstract

PURPOSE:To improve the compression ratio and to decrease deterioration in picture quality by not applying compression to a consecutive frame picture at a constant compression ratio but applying compression in the unit of frames based on a characteristic of the picture. CONSTITUTION:An input picture is converted into digital picture data at an input section 1, converted into a matrix sequentially at a matrix conversion section 2, decomposed into a frequency component by DCT arithmetic operation at a DCT(discrete cosine transformation) arithmetic operation section 3 and the DCT component is stored tentatively in a storage section 4 and the DC component is outputted to a DC component storage section 5 and a difference arithmetic operation section 6. The arithmetic operation section 6 uses the current frame DC component and the DC component of the same coordinate of one preceding frame stored in the storage section 5 to implement difference arithmetic operation. A compression parameter setting section 7 decides the compression parameter in response to the size of the arithmetic operation result. The DCT component stored tentatively in the storage section 4 is quantized at a quantization section 8 by using the compression parameter and further compressed by a Huffman transformation section 9.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image compression apparatus suitable for compressing image data such as a moving picture on a TV.

[0002]

2. Description of the Related Art Generally, in a system that handles moving images, a large amount of data is a problem from the viewpoint of storage capacity, transfer capacity of a bus, processing capacity of a processor, etc. However, the amount of data was reduced. However, if the compression rate is simply increased to reduce the amount of data, the amount of data required to reproduce the image will be reduced, and the image quality will be greatly degraded.

[0003]

As described above, in a system that handles moving images, the compression rate has been increased to reduce the amount of data. However, simply increasing the compression rate only deteriorates the image quality. Therefore, there has been a demand for an efficient compression technique that reduces the image quality and increases the compression rate.

The present invention has been made to solve such problems, and a first object thereof is to use a method of compressing a continuous frame image at a uniform compression ratio without It is an object of the present invention to provide an efficient image compressing device that reduces the image quality and increases the compression rate by performing the compression on a frame-by-frame basis.

A second object of the present invention is not to handle continuous frame images at a constant frame rate, but to
An object of the present invention is to provide an efficient image compression apparatus that accurately grasps a change in an image and changes a frame rate based on the characteristics of the image to reduce the total amount of information, reduce image quality degradation, and increase a compression rate.

[0006]

The first structure of the present invention is as follows.
The input means for inputting the moving image and the input moving image input by this input means are sequentially subjected to DCT calculation and D
DCT calculation means for outputting a CT component, storage means for temporarily storing the DCT component output by the DCT calculation means, and the DC temporarily stored by the storage means
An image change extraction means for performing a difference operation between the T component and the DCT component of the next image output by the DCT operation means, and extracting an image change according to the operation result, and a change extracted by the image change extraction means. A compression parameter setting means for setting a compression parameter according to the amount and a compression / expansion means for compressing or expanding the DCT component temporarily stored with the compression parameter set by the compression parameter setting means are provided. And

Further, the compression parameter setting means is a frame in which the change in DCT component is larger than that in the previous frame,
That is, the compression rate should be set to 1 or a low value in a frame in which there is a scene change or a large motion, and the compression rate should be set to a high value in a frame in which a change in the DCT component is small, that is, an image in which the motion is small and close to the previous frame. The image compression rate is changed according to the amount of change from the previous frame.

According to a second aspect of the present invention, input means for inputting a moving image, DCT operation means for sequentially performing DCT operation on the moving image input by this input means and outputting a DCT component, and output by this DCT operation means. DC
A storage unit that temporarily stores the T component, a difference calculation is performed between the DCT component that is temporarily stored by the storage unit and the DCT component of the next image that is output by the DCT calculation unit, and the result of the calculation changes the image. Image change extraction means for extracting the amount, frame rate setting means for setting a frame rate which is the number of frames per unit time according to the change amount extracted by the image change extraction means, and frame at the set frame rate And a frame number changing means for changing the number.

Further, the frame rate setting means is a frame in which the change of the DCT component is larger than that of the previous frame,
That is, the frame rate is set to a high value in a frame in which there is a scene change or a large motion, and the frame rate is set to a low value in a frame in which a change in the DCT component is small, that is, an image in which the motion is small and close to the previous frame. It is characterized in that the frame rate is changed according to the amount of change from the previous frame.

Further, the image change extraction means is a DCT.
The difference between the DC component and the AC component of each component is calculated for each frequency component, and the amount of change in the image is extracted based on the calculation result.

The number-of-frames changing means temporarily stores the DCT component output by the DCT calculating means, and controls the output of the stored DCT component according to the frame rate.

[0012]

In the first structure described above, the DCT calculation is sequentially performed on the moving image in frame units, and in a frame in which the change in DCT component is larger than that in the previous frame, that is, a frame in which a scene change or a large motion occurs, -By compressing with a low compression rate, the image that is as close to the original image as possible is left. On the other hand, for a frame in which the change in DCT component is small, that is, an image in which the motion is small and close to the previous frame, the compression rate is set to a high value for compression. In this way, the overall compression rate is increased by changing the compression rate according to the amount of change from the previous frame.

In the above second structure, for moving images,
A DCT operation is sequentially performed on a frame-by-frame basis, and a frame having a large difference from the DCT component located at the same coordinate of the previous frame,
That is, in a frame in which there is a scene change or a large movement, the transfer rate (frame rate) is set to a high value and the transfer is performed to leave an image as close to the original image as possible. on the other hand,
For a frame in which the change in DCT component is small, that is, an image in which the motion is small and close to the previous frame, the frame rate is set to a low value for transfer. By thus changing the frame rate according to the amount of change in the image, the overall compression rate is increased.

Further, in the first and second configurations, since the DCT component that appears in the process of JPEG compression is used as the image change detection means, when it is applied to the image compression apparatus of the JPEG compression method. , DCT arithmetic circuit JP
It can be commonly used for EG compression and image change detection,
The circuit configuration is simplified and the circuit scale is reduced as compared with the case where another image change detection method is adopted.

[0015]

Embodiments of the present invention will be described below with reference to the drawings.

First, a case where the first embodiment of the present invention is applied to a JPEG image compression apparatus will be described. JPEG is a committee (Joint Photographic Expert Grou) that conducts international standardization work for image compression methods for color still image transmission and storage by jointly using CCITT and ISO.
p) and the still image compression method defined here is called the JPEG method. Figure 1
It is a schematic block diagram of the image compression apparatus in the same Example. In the image compression apparatus shown in FIG.
This embodiment will be described by taking the case of using an image of 0 × 480, 30 fps (frames / second) as an example.

In the figure, the input unit 1 converts an input image into 64
This is a circuit for converting to 0 × 480, 30 fps (non-interlace). For example, if the input signal is an analog component signal, sampling is performed at about 12 MHz. The input section 1 includes a matrix conversion section 2 and a DC.
The T calculation unit 3 is connected. The matrix conversion unit 2 is
The line signal is converted into an 8 × 8 matrix for DCT calculation, and is connected to the DCT calculation unit 3.

The DCT calculation unit 3 sequentially performs DCT calculation on the image matrix-converted by the matrix conversion unit 2, and is connected to the storage unit 4, the DC component storage unit 5 and the difference calculation unit 6. The calculation unit 3 decomposes the input signal into frequency components by DCT calculation and outputs the frequency components to the storage unit 4 as DCT components. Further, the calculation unit 3 outputs only the DC component in the DCT component to the DC component storage unit 5 and the difference calculation unit 6.

The storage unit 4 temporarily stores the DCT component output from the DCT calculation unit 3, and is connected to the quantization unit 8. The storage unit 4 sets the temporarily stored DCT component to 1
It is delayed by the frame and output to the quantizer 8.

The DC component storage unit 5 temporarily stores the DC component output from the DCT calculation unit 3, and is connected to the difference calculation unit 6. The storage unit 5 stores the temporarily stored DC
The component is delayed by one frame and output to the difference calculation unit 6.

The difference calculation unit 6 is a circuit for calculating the difference between the DC components located at the same coordinates of the preceding and following frames, and is connected to the compression parameter setting unit 7. Same computing unit 6
Is a DC component sequentially sent from the DCT calculation unit 3,
The difference between the DC component stored in the DC component storage unit 5 which is the addition data storage unit and located at the same coordinate one frame before is calculated to calculate the change amount of the DC component, and the compression parameter setting unit. Output to 7.

The compression parameter setting unit 7 includes a difference calculation unit 6
This is a circuit that sets the compression parameter according to the amount of change in the DC component calculated in step. The setting unit 5 is connected to the quantizing unit 8 and an external device (not shown), selects an appropriate compression parameter from among several compression parameters prepared in advance, and selects the quantization parameter of the quantizing unit 8. It is set in a conversion table (not shown) and is output to an external device (not shown). Information about which quantization table is selected (compression ratio information) is provided when the compression data compressed by the compression parameter set in the quantization table is output to an external device (not shown). , To the external device (not shown) together.

The quantizer 8 outputs the D output from the memory 4.
The CT component is quantized with the compression parameter set by the compression parameter setting unit 7. The quantizing unit 8 is connected to the Huffman transforming unit 9 to quantize the quantized result.
It is output to the conversion unit 9 as a data.

The Huffman conversion unit 9 is a circuit for Huffman coding to further compress the quantized data. The converter 9 is connected to an external device (not shown),
The quantized data is further compressed and this compressed data is output to an external device (not shown). Next, the operation of the configuration of FIG. 1 will be described.

First, the input image is sampled by the input unit 1 and becomes digital image data of a predetermined size. This data is converted to DCT by the matrix conversion unit 2.
Sequential conversion to an 8 × 8 matrix, which is the unit of calculation,
It is sent to the DCT calculation unit 3. The DCT calculation unit 3 performs DCT calculation on the data matrix-converted by the matrix conversion unit 2, and stores the DCT data (DCT data) in the storage unit 4.
Component), and outputs only the DC component in the DCT component to the DC component storage unit 5 and the difference calculation unit 6.

The storage unit 4 always delays the data (DCT component) transferred from the DCT calculation unit 3 by one frame,
Output to the quantizer 8. This is because the compression parameter is set by calculation between frames, and when the setting of the compression parameter is completed, the DCT component of the next frame has already been output, and a time shift occurs between the compression parameter and the DCT component. is there.

The difference calculation unit 6 sequentially reads the DC component of the current frame from the DCT calculation unit 3, and also reads the DC component of the same coordinate of the frame one screen before from the DC component storage unit 5 to calculate the difference. To do. After that, the calculation unit 6
The difference data as the result of this calculation is summed up in frame units and transferred to the compression parameter setting unit 7.

If the calculation result (difference data) is a large value, there is a scene change or a large image change, and if it is small, it can be determined that there is almost no image change.
Since the compression parameter setting unit 7 determines the magnitude of the calculation result, a threshold is set in advance in the setting parameter setting unit 7, and the difference that is the calculation result sent to the setting unit 7 based on that threshold value is set. It suffices to judge whether the data is large or small.

As described above, it is an existing technique to extract the change of the image according to the change amount of the DCT component between frames and use it for the detection of the scene, and there are some methods other than this method. Therefore, any method may be used as long as it is a technique capable of extracting the amount of change.

The compression parameter setting section 7 includes a difference calculating section 6
In accordance with the size of the difference data transferred from, the appropriate one is selected from among several prepared quantization parameters (compression parameters) and set in the quantization table of the quantization unit 8. .

The image compression rate is determined by this quantization parameter. For example, when dividing the size of the difference data into n stages, there must be n sets of quantization parameters, and depending on the size of the difference data, one of the n sets can be used.
A set is selected and set in the quantization table.
In this embodiment, the quantization parameter is set to a small value when the difference data is large, and the quantization parameter is set to a large value when the difference data is small.

The quantization parameter set in the quantization table of the quantization unit 8 is the DCT output from the storage unit 4.
It has the same size of 8 × 8 (64) as the component. The quantizing unit 8 divides the 8 × 8 DCT component by the value of the corresponding quantization table (quantization parameter set by the compression parameter setting unit 7), and quantizes the DCT component. To do. The larger the value of the quantization table, the smaller the divided value. Therefore, less data is required for image representation, and the compression rate increases. In the present embodiment, as described above, when the difference data is large, the quantization parameter is set to a small value, and when the difference data is small, the quantization parameter is set to a large value. Therefore, as a result of the quantization of the DCT component by the quantizer 8, the compression rate is low when the difference data is large, and the compression rate is high when the difference data is small. The quantizer 8 quantizes the result of quantizing the DCT component.
It is output to the Huffman conversion unit 9 as a data.

The Huffman conversion unit 9 converts the quantized data into Huffman codes for further compression, and outputs the compressed data compressed by the Huffman coding to an external device (not shown).

As described in detail above, when there is a scene change or a large image change, the compression rate is low and the image is close to the original image. On the other hand, when the image is the same as the previous frame or the image change is small, compression is performed. The rate is high. As described above, in this embodiment, the compression rate can be changed according to the change amount of the image. Next, a case where the second embodiment of the present invention is applied to an image compression device will be described. FIG. 2 is a schematic configuration diagram of the image compression apparatus in the embodiment.

Hereinafter, in the image compression apparatus shown in FIG.
Image size 640 × 480, 30 fps as in the first embodiment
The present embodiment will be described by taking the case of using an image of (frame / second) as an example.

In the figure, an input unit 11 is a circuit for converting an input image into 640 × 480, 30 fps (non-interlace), as in the first embodiment.
It is a circuit similar to. A matrix conversion unit 12 and a DCT calculation unit 13 are connected to the input unit 11.
The matrix conversion unit 12 converts a line signal for DCT calculation into an 8 × 8 matrix, and is connected to the DCT calculation unit 13 and the storage unit 14.

The DCT calculator 13 is the matrix converter 1
The DCT calculation is sequentially performed on the image matrix-converted in 2 and is connected to the DC component storage unit 15 and the difference calculation unit 16. The calculation unit 13 outputs the DC component in the DCT component, which is the calculation result, to the DC component storage unit 15 and the difference calculation unit 16.

The storage unit 14 temporarily stores the image data matrix-converted by the matrix conversion unit 12, and is connected to an external device (not shown). The storage unit 14 uses the frame rate parameter output from the rate setting unit 17 connected to the storage unit 14 to store the corresponding data among the stored image data in an external device (not shown). Output to.

The DC component storage unit 15 stores the DC component of the newly input frame (D of the new frame until the instruction from the difference calculation unit 16 connected to the storage unit 15).
(C component) is temporarily stored. In addition, the same storage unit 15
Stores the DC component of the previous frame (DC component of the old frame) for one frame, and outputs the DC component of the old frame to the same computing unit 16 under the control of the difference computing unit 16. Thereafter, the storage unit 15 updates the DC component of the old frame with the DC component of the new frame or leaves the DC component of the old frame as it is, according to an instruction from the rate setting unit 17 connected to the storage unit 15. , Erases the DC component of a new frame.

The difference calculator 16 is a circuit for calculating the difference between the DC component of the new frame sequentially sent from the DCT calculator 13 and the DC component of the old frame sent from the DC component storage 15. The calculation unit 16 is connected to the rate setting unit 17, calculates the difference between the DC component of the new frame and the DC component of the old frame, calculates the change amount of the DC component, and outputs the change amount to the rate setting unit 17. To do.

The rate setting section 17 is a circuit for setting a frame rate parameter. The setting unit 17 includes a storage unit 14, a DC component storage unit 15 and an external device (not shown).
The frame rate parameter is set according to the change amount of the DC component calculated by the difference calculation unit 16, and is output to the storage unit 14, the DC component storage unit 15 and an external device (not shown). Next, the operation of the configuration of FIG. 2 will be described. First, the input image is sampled by the input unit 11 and becomes digital image data of a predetermined size. This data is converted to DCT by the matrix conversion unit 12.
Sequential conversion to an 8 × 8 matrix, which is the unit of calculation,
It is sent to the storage unit 14 and the DCT calculation unit 13.

The DCT calculator 13 is the matrix converter 1
The DCT operation is performed on the data matrix-converted by 2, and the DC component in the DCT component is stored in the DC component storage unit 1.
5 and the difference calculation unit 16.

The storage unit 14 temporarily stores the image data converted by the matrix conversion unit 12, and the rate setting unit 1
According to the instruction of 7, the stored image data may or may not be output to an external device (not shown). When outputting image data to an external device (not shown), the storage unit 14 delays the output by about one frame and outputs the delayed image data. This is because the frame rate parameter is set by calculation between frames, and when the setting of the frame rate parameter is completed, the image data of the next frame is already output, and the output of the frame rate parameter and the image data are performed.
This is because there is a time lag in the output of the data.

The difference calculation unit 16 calculates the difference between the DC component of the new frame sequentially sent from the DCT calculation unit 13 and the DC component of the old frame stored in the DC component storage unit 15 and located at the same coordinate. Then, the resulting difference data is summed up for each component and transferred to the rate setting unit 17. If this calculation result (difference data) has a large value, there is a scene change or a large image change, and if it is small, it can be determined that there was almost no image change.

The rate setting unit 17 determines a frame rate parameter for instructing whether or not to thin out the image (new frame) according to the size of the difference data which is the transferred difference calculation result. , The same parameter storage unit 14, DC component storage unit 15 and an external device (not shown)
Output to.

In response to this, the DC component storage unit 15
When thinning out an image, the DC component of the old frame stored in the storage unit 15 is stored as it is, the DC component of the new frame sent from the DCT calculation unit 13 is erased, and a new image is not thinned out. When adopting a frame, the DC component of the new frame is updated with the DC component of the new frame.

As the number of images to be thinned out increases, the amount of data required to express an image per unit time decreases, so that the frame rate (the number of frames per unit time) decreases and the data compression rate increases, but the image quality Will get worse. Conversely, the smaller the number of images to be thinned out, the higher the frame rate and the lower the data compression rate, but the high image quality can be maintained.

In the rate setting unit 17, a threshold is set in advance as a criterion for determining the magnitude of difference data for determining the frame rate parameter. The rate setting unit 17 sets the frame rate parameter to “0” if the difference data from the difference calculation unit 16 satisfies the respective thresholds (if smaller than the thresholds), and if not satisfied. Set the frame rate parameter to "1" (if greater than the threshold). The larger the threshold value is, the lower the frame rate is, and the higher the compression rate is. On the contrary, the smaller the value is, the closer the frame rate is to the original frame rate, and the closer to the original image. The frame rate varies greatly depending on how this threshold is set, so it is necessary to set it to an appropriate value.

The storage unit 14 receives the frame rate parameter output from the rate setting unit 17, and outputs nothing to an external device (not shown) when the parameter is "1", that is, when thinning out images. When the parameter is "0", that is, when the images are not thinned, the image data transferred from the matrix conversion unit 12 and temporarily stored is output to an external device (not shown). As a result, the storage unit 1
The image data from No. 4 is compressed data in which the total number of frames is reduced according to the frame rate parameter.

When reproducing these data,
The frame image is temporarily stored, and if the frame rate parameter is "1", new frame data is output,
If it is "0", old frame data may be output.

When the data compressed by these means is reproduced, a new frame is displayed when the image changes more than the value set at the time of compression, and a new frame is displayed when the change does not exceed a certain value. The same image is displayed until. That is, when the change in the image is small, the same image is displayed as it is, and when the change in the image is large, a new frame is displayed. As described above in detail, in the present embodiment, efficient image compression can be performed by changing the frame rate according to the degree of image change. Next,
A case where the third embodiment of the present invention is applied to a JPEG image compression apparatus will be described. FIG. 3 is a schematic configuration diagram of the image compression apparatus in the embodiment.

As shown in the figure, the image compression apparatus according to the present embodiment includes an input unit 21, a matrix conversion unit 22, a DCT calculation unit 23, a storage unit 24, a DCT component storage unit 25, a difference calculation unit 26, and a rate setting unit 27. , A quantization unit 28 and a Huffman conversion unit 29.

In the figure, an input unit 21, a matrix conversion unit 22, a quantization unit 28 and a Huffman conversion unit 29 are the same as the input unit 1, matrix conversion unit 2, quantization unit 8 and Huffman conversion unit 9 of FIG. 1, respectively. However, the description thereof will be omitted.

In the above structure, the DCT calculator 23
Similar to the first embodiment, the DCT calculation is sequentially performed on the image matrix-converted by the matrix conversion unit 22, and is connected to the storage unit 24, the DCT component storage unit 25, and the difference calculation unit 26. The calculation unit 23 outputs the DCT component to the storage unit 24, the DCT component storage unit 25, and the difference calculation unit 26 as in the first embodiment.

However, in the first embodiment, the DCT calculator 3
Is the DCT component storage unit 5
And the difference was output to the difference calculation unit 6, but in the present embodiment,
The DCT calculator 23 outputs the AC component as well as the DC component.

Further, in the second embodiment, the matrix-converted image data is directly output to the storage unit 4, but in the present embodiment, since the JPEG compression method is adopted, D
The storage unit 2 stores the DCT component obtained by CT calculation as image data.
It is output to 4. Hereinafter, D used as image data
In order to distinguish from the CT component, the DCT component used for setting the frame rate parameter is called the DC or AC component.

The storage unit 24 temporarily stores the DCT component output from the DCT calculation unit 23, and is connected to the quantization unit 28. The storage unit 24 is the same as the storage unit 24.
The corresponding information (DCT component) among the stored information is output to the quantization unit 28 according to the frame rate parameter output from the rate setting unit 27 connected to the.

The DCT component storage unit 25 temporarily stores the DC and AC components of the newly input frame (DC and AC components of the new frame) until an instruction is given from the difference calculation unit 26 connected to the storage unit 25. Remember. In addition, the storage unit 25 stores the DC and AC components of the previous frame (DC and AC components of the old frame) for one frame, and the DC and AC components of the old frame are controlled by the difference calculation unit 26. Is output to the calculation unit 26. Thereafter, the storage unit 25 updates the DC and AC components of the old frame with the DC and AC components of the new frame, or updates the DC of the old frame according to an instruction from the rate setting unit 27 connected to the storage unit 25. , AC components are left as they are, and DC and AC components of a new frame are erased.

The difference calculator 26 receives the DC and AC components of the new frame sequentially sent from the DCT calculator 23 and D.
D of the old frame sent from the CT component storage unit 25
This is a circuit that calculates the difference between the C and AC components. The calculation unit 26 is connected to the rate setting unit 27, calculates the difference between the DC and AC components of the new frame and the DC and AC components of the old frame, and calculates the change amounts of the DC and AC components, Output to the rate setting unit 27.

The rate setting section 27 is a circuit for setting a frame rate parameter. The setting unit 27 is connected to the storage unit 24, the DCT component storage unit 25, and an external device (not shown), and D calculated by the difference calculation unit 26.
A frame rate parameter is set according to the amount of change in the C and AC components, and is output to the storage unit 24, the DCT component storage unit 25, and an external device (not shown).

In the rate setting unit 27, thresholds are preset for each frequency as a criterion for determining the magnitude of the difference data for determining the frame rate parameter. The rate setting unit 27 sets the frame rate parameter to “0” if the difference data for each frequency is smaller than each threshold, and sets the frame rate parameter to “1” if it is larger than the threshold. The larger the threshold value is, the lower the frame rate is, and the higher the compression rate is. On the contrary, the smaller the value is, the closer the frame rate is to the original frame rate, and the closer to the original image.

Even if the actual image does not change so much as a whole, there are many cases where changes occur in small parts. In the present embodiment, not only the DC component but also the AC component is used for detecting the image change, so that the minute image change is also detected. For this reason, when looking at the changes in the detected image by frequency component, the amount of change in the high-frequency component becomes larger than the amount of change in the low-frequency component, so the threshold of the high-frequency component must be set to a high value to some extent. . The image change detection method in this embodiment can be applied to the first embodiment.

Since the frame rate greatly changes depending on how the threshold is set, it needs to be set to an appropriate value. In some cases,
The threshold may be variable depending on the distribution of frequency components.

The storage unit 24 receives the frame rate parameter output from the rate setting unit 27, outputs nothing to the quantization unit 28 when the parameter is "1", that is, when thinning out an image, and the same parameter is output. Is "0", that is, when the image is not thinned, the quantizing unit 28 uses the DCT component transferred from the DCT calculating unit 23 and temporarily stored.
Output to.

The quantizing unit 28 quantizes the DCT component from the storage unit 24 with the set compression parameter, and outputs the result to the Huffman transform unit 29 as quantized data. The Huffman conversion unit 29 converts the quantized data sent from the quantization unit 28 into Huffman code for further compression.

The compression data and the frame rate parameter after the Huffman conversion are output to an external device (not shown).
Further, when reproducing these data, like the second embodiment, the frame image is temporarily stored, and if the frame rate parameter is "1", new frame data is output,
If it is "0", old frame data may be output.

In this embodiment, the DC used in JPEG compression for detecting the change in the image for changing the frame rate.
T calculation is used, and the result of this DCT calculation is JPEG
It can be commonly used for compression and image change detection. Therefore, in the present embodiment, efficient compression can be performed by using an image compression technique such as JPEG compression together.
The circuit configuration can also be simplified.

[0068]

According to the present invention, a DCT operation is performed on a moving image on a frame-by-frame basis, and the DCT calculation is performed in comparison with the previous frame.
The compression rate of the image data is set to a low value for the frame in which the change of the component is large, and the compression rate of the image data is set for the frame in which the change of the DCT component is small. Since the compression rate of the image of the current frame is changed according to the amount of change from the previous frame, such as setting a high value for compression, the overall compression rate is improved. In other words, a scene transition or a portion with a lot of movement, that is, a portion that has a strong impact on the viewer, has a relatively large amount of image information and is close to the original image, and a portion with little movement has a compression degree. Since the image quality is high and the amount of information is small, the deterioration of the apparent image quality is reduced as a whole, and the degree of compression is also improved. With this, compared to the case of simply increasing the compression ratio and compressing,
Images obtained with the same amount of information have more natural and high image quality.

When an image is searched by adding an index to a scene change, a compressed image with poor image quality is displayed if the compression ratio is uniform as in the conventional case. According to this, images having a low compression rate and a large amount of information are continuously displayed. Also, when used as a still image, an image close to the original image can be obtained by extracting the scene transition.

Further, according to the present invention, for moving images,
The DCT calculation is sequentially performed on a frame-by-frame basis, and for a frame having a large difference from the DCT component located at the same coordinate of the previous frame, the transfer rate (frame rate) is set to a high value and the transfer is performed. For frames with small changes in components, the frame rate is changed according to the amount of change in the image, such as setting the frame rate to a low value and transferring, so the overall compression rate is improved. To do. In other words, the image with a large amount of information is left in the form of a large amount of information close to the original image, and the part with little movement is reduced in frame rate to reduce the amount of information. The compression degree of is also improved. As a result, an image obtained with the same amount of information has a more natural and high image quality as compared with a case where the frame rate is simply reduced and compression is performed.

Further, according to the present invention, since the DCT component that appears in the process of JPEG compression is used as the image change detection means, the DCT arithmetic circuit can be used commonly for JPEG compression and image change detection. it can. For this reason, the circuit configuration of the device is simplified and the circuit scale is smaller than in the case where the image change detection is performed by another method. Moreover, instead of using only the DC component of the DCT component as in the past,
Since the AC component is also subjected to difference calculation for each frequency component and is used for image change extraction, it is possible to more accurately extract the image change even in response to fine movement of the image.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of a JPEG image compression apparatus according to a first embodiment of the present invention.

FIG. 2 is a schematic configuration diagram of an image compression apparatus according to a second embodiment of the present invention.

FIG. 3 is a schematic configuration diagram of a JPEG image compression apparatus according to a third embodiment of the present invention.

[Explanation of symbols]

1, 11, 21 ... Input unit, 2, 12, 22 ... Matrix conversion unit, 3, 13, 23 ... DCT calculation unit, 4, 24 ...
(DCT component) storage unit, 14 ... (Image data) storage unit,
5, 15 ... DC component storage unit, 25 ... DCT component storage unit,
6, 16, 26 ... Difference calculation unit, 7 ... Compression parameter setting unit, 17, 27 ... Rate setting unit, 8, 28 ... Quantization unit,
9, 29 ... Huffman conversion unit.

Claims (6)

[Claims] 1. An input device for inputting a moving image, and DC for the moving images input by the input device.
T (Discrete Cosine Transform)
DCT computing means for performing computation and outputting a DCT component, storage means for temporarily storing the DCT component output by the DCT computing means, and the DCT component temporarily stored by the storage means and output by the DCT computing means DC of the next image to be rendered
An image change extraction unit that performs a difference calculation with the T component and extracts the amount of change in the image based on the calculation result, and a compression parameter that determines the compression ratio of the image according to the amount of change extracted by the image change extraction unit An image compression apparatus comprising: a compression parameter setting means to be set; and an image compression / decompression means for compressing or expanding an image according to the compression parameter set by the compression parameter setting means. 2. The compression parameter setting means sets the compression ratio to 1 or a low value in a frame in which the change in the DCT component is large compared to the previous frame, and sets a high compression ratio in a frame in which the change in the DCT component is small. The image compression apparatus according to claim 1, wherein the image compression apparatus is set to a value, and the compression rate is changed according to the change amount from the previous frame. 3. Input means for inputting a moving image and DC for moving images input by this input means
DCT calculation means for performing T calculation and outputting a DCT component, storage means for temporarily storing the DCT component output by the DCT calculation means, and the DCT component and the DCT calculation means temporarily stored by the storage means DC of the next image to be output
An image change extraction unit that performs a difference calculation with the T component and extracts the amount of change in the image based on the calculation result, and a frame rate that is the number of frames per unit time according to the change amount extracted by the image change extraction unit. An image compression apparatus comprising: a frame rate setting unit for setting the number of frames; and a frame number changing unit for changing the number of frames at the frame rate set by the frame rate setting unit. 4. The frame rate setting means sets a frame rate to a high value in a frame in which the change in the DCT component is large compared to the previous frame, and sets a low frame rate in a frame in which the change in the DCT component is small. The image compression apparatus according to claim 3, wherein the frame rate is set to a value and the frame rate is changed according to the change amount from the previous frame. 5. The image change extraction means is configured to perform a difference calculation for each of the DC component and the AC component of the DCT component for each frequency component, and extract the amount of change in the image based on the result of the difference calculation. The image compression device according to claim 3. 6. The frame number changing means is the DCT.
Temporarily storing the DCT component output by the computing means,
The image compression apparatus according to claim 3, wherein the output of the stored DCT component is controlled according to the frame rate.