JP2001346211A - Television receiver, image magnification device, processing method for television signal, image magnification method and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce ringing caused around an edge when applying magnification conversion of number of pixels in a DCT area to an image including many edges and to enhance the image quality after the magnification. SOLUTION: Applying mirror image reflection to a received original signal in a frequency direction by using a Nyquist frequency before re-sampling the frequency component of the received original signal generates a 2nd signal component, applying mirror image inversion to the 2nd signal component generates a 3rd signal component, and adding the 3rd signal component to the original signal as its high frequency component stimulatingly generates its high frequency component. Applying filter processing to the frequency components from the low frequency component of the 1st signal component to the Nyquist frequency band after the re-sampling can magnify video contents including many edges with high quality.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a television receiver, an image enlargement apparatus, a method for processing a television signal,
More particularly, the present invention relates to an image enlargement method and a storage medium, in particular, a television receiver and an image enlargement device, more specifically, a digital television receiver that receives a transport stream encoded by MPEG2, decodes the transport stream, and displays it. In the input / output device in the time space domain,
The present invention relates to a television receiver capable of enlarging and converting an input signal to a resolution of a display device or an arbitrary resolution for display.

[0002]

2. Description of the Related Art With the recent development of digital technology, the television broadcasting system has been changed to analog NTSC / PA.
The L / SECAM system is shifting to a digital system. For example, for terrestrial digital broadcasting, Japan, the United States,
Different broadcast systems have been developed or are already in operation in Europe.

For example, in the United States, the development of digital terrestrial broadcasting ATV (Advanced TV) has been progressing since 1987, and the system standard was established in 1995 in accordance with ATSC (Advanced TV).
Television Systems committee) from FCC (Federa
l communications commission), and was determined in 1996 except for the standards such as the number of effective scanning lines and the number of horizontal pixels. In Europe, DVB-T (Digi
tal Video Broadcasting-Terrestrial) has been determined as a common European system.

The major difference between the two systems is the modulation system. The U.S.A. system is called an 8-level VSB (Vestigial Side Bands) system, uses a single carrier, divides the amplitude into eight stages, and This is a method of transmitting a signal by magnitude.

The European system is based on OFDM (Orthogonal Freq).
This method is called a uency division multiplexing method, and uses a few hundred carriers to reduce the signal transmission speed per wave compared to the case of a single carrier, thereby improving the resistance to ghost interference. In both the United States and Europe, the target is fixed reception, and reception by mobiles is difficult.

[0006] In Japan, discussions are currently underway for standardization, but studies are being made on a system that can receive not only fixed reception but also a mobile body. In particular, NHK (Japan Broadcasting Corporation) has BST (Band Segmented Transmission)
-We are studying a new digital modulation transmission method called FDM.

In the BST-OFDM, a plurality of types of modulation schemes can be designated in one band, whereby a broadcasting station can use a plurality of modulation schemes such as high-vision for fixed reception, images for mobile reception, and data services. This is a system that enables various types of overnight data transmission services.

On the other hand, various formats are assumed for the number of horizontal and vertical pixels of the digital broadcast. For example, as HD (High definition) so-called high-definition, 1920 × 1080I, 1440 × 1080
I, 1280 × 720P, 720 × 480P (horizontal × vertical, I indicates interlaced scanning, P indicates sequential scanning), SD (Stan
720 × 480I, 54 as the dard definition)
4 × 480I, 480 × 480I and the like.

At present, ARIB (radio wave industry) standardizes these image formats as standards for the 2000 BS digital broadcasting, and a draft was issued in October 1999. In this draft, the horizontal of the stream (input signal) and the display device (output signal),
It also describes the standard for conversion to the number of pixels when the number of vertical pixels is different.

For example, the above 720 × 480I (aspect ratio 4: 3) SD signal is converted into a 1920 × 10801 HD signal.
When displaying on a display device (16: 9), 9
/ 4 times and doubling in the horizontal direction to add 480 side panels (black borders) (see FIG. 12A).

Also, for example, a letter box format 720
× 480I (360I) HD signal is 1920 × 108
When the image is displayed on the HD display device (16: 9) of 0I, the image is tripled in the vertical direction and 8/3 in the horizontal direction.

As the conversion method of the number of pixels, there are a nearest neighbor method (Nearest Neighbor) and a linear interpolation method (Bi-Linea).
r) A cubic convolution interpolation method (Cubic Convolution) or the like is generally used.

[0013] In these systems, FIR (Finite Impulse) is obtained from discrete sampling data in the time-space domain.
Response) A method in which a continuous envelope is reproduced using a filter, and an interpolated pixel is found on this envelope.

In the method using these FIR filters, the key is how to use an FIR filter having an ideal low-pass characteristic. Generally, the number of taps of the filter is increased, and a filter having a characteristic of f = Sin (πx) / (πx), which has a characteristic close to a so-called Sinc function, is approximated by a finite length and used.

Of course, the multi-tap can provide characteristics closer to the Sinc function, but the cost increases accordingly. Furthermore, instead of the above space-time domain, MPEG
(Moving Picture Expert Group) DCT (Discreet
A method of converting the number of pixels in a Cosine Transform area has been proposed (for example, Japanese Patent Laid-Open No. 6-233271).
issue).

The conversion of the number of pixels in the DCT area will be described with reference to FIGS. FIG. 3 is a block diagram showing a configuration of a general MPEG decoder.
1 is a buffer means, 32 is a variable length decoding means, 33 is an inverse quantization means, 34 is an inverse DCT means, 35 is a motion compensation prediction means, 36 is a video memory means, and 37 is a format conversion means.

As shown in FIG. 3, the decoder means performs variable length decoding, inverse quantization,
Inverse DCT processing and motion compensation are performed to decode the signal before encoding.

The conversion of the number of pixels in the DCT area is as follows.
Immediately before the inverse DCT means 34 in FIG.
3).

FIG. 4 is a conceptual diagram for explaining the enlarging process in the DCT area of the conventional example, and FIG.
A DCT coefficient configured in units of eight pixels is shown, and the lower right of the DCT coefficient becomes a high-frequency component.

Therefore, for example, as shown in FIG. 4B, “0 value” is additionally inserted into the high frequency components on the right side and lower side (gray area) of the 8 × 8 DCT coefficient, and 10 ×
The inverse DCT is performed with the 10 base matrices, and the gain is adjusted (10/8 times) in accordance with the enlargement ratio to obtain 10/8.
Double magnification processing can be realized.

[0021]

Since 2000, in order to spur the transition from analog broadcasting to digital broadcasting, programs with higher image quality, that is, higher definition, and containing more high-frequency components (edges) in spatial frequency have been proposed. (Images), the need to distribute content is increasing.

The above-described enlargement conversion of the number of pixels in the DCT area is a suitable method in a system for decoding an encoded digital broadcast of the MPEG2 system, and the compression ratio is increased at the expense of a high frequency component in a spatial frequency. Although it is effective for the current natural moving image, when applied to an image including many high-frequency monuments (edges) at the spatial frequency as described above, a lot of ringing occurs around the edge, There is a problem that the image quality is significantly deteriorated.

In view of the above-described problems, the present invention reduces ringing generated around an edge when an image including many edges is subjected to enlargement conversion of the number of pixels in a DCT region, and includes an edge. An object of the present invention is to improve the image quality after enlargement for an area that does not exist.

[0024]

SUMMARY OF THE INVENTION A television receiver according to the present invention has a mirror image folded in the frequency direction around a Nyquist frequency before resampling a first signal component which is a frequency component of an input original signal. After performing DCT
Folding and inverting means for generating a second signal component shifted in the high frequency direction by one component of the coefficient, inverting means for mirror-inverting the second signal component to generate a third signal generation, and the inverting means High-frequency component generation means for adding the third signal component generated by the above as a high-frequency component of the input original signal to pseudo-generate a high-frequency component;
And a filter means for filtering a frequency component from the low-frequency component of the signal component to the Nyquist frequency band after the resampling.
Another feature of the present invention is that a pseudo high frequency component up to the Nyquist frequency after sub-sampling the first signal component, which is a frequency component of the input original signal,
Insertion means for additionally inserting the high frequency side of the original signal; and filter means for filtering frequency components from a low frequency component of the first signal component to a pseudo high frequency component inserted by the insertion means. It is characterized by having.
Another feature of the present invention is that a variable length decoding means for performing variable length decoding on the encoded bit stream, and a quantized DCT coefficient of a predetermined block decoded by the variable length decoding means. Inversely quantizing means for inversely quantizing, and the quantization D restored by the inversely quantizing means.
Edge detection means for detecting an edge in block units using CT coefficients, a first block detected to include an edge by the edge detection means, and a different enlargement for a second block detected not to include an edge. Image processing means for performing processing, and inverse DCT for performing inverse DCT on a DCT coefficient block after image processing by the image processing means using a base matrix corresponding to the enlarged block size
T means, local decoding by motion compensation in consideration of the conversion of the number of pixels, decoding means for decoding the bit stream into a video signal before being encoded, and A decoder having pixel number conversion means for converting the number of pixels to an appropriate number of pixels to be displayed on the display means; and an image resolution of a video signal output from the decoder means, based on information including edge information. ,
System control means for controlling the operation of the decoder means, and display control means for converting a signal output from the decoder means into a video signal displayable on the display means and controlling the display means. Features. According to another feature of the present invention, the image processing unit includes a first conversion processing unit that performs a pixel number conversion process on a block detected to include the edge, and does not include the edge. Second conversion processing means for performing a pixel number conversion process on the detected block, wherein the first conversion processing means includes a gain adjustment means for adjusting a gain according to an enlargement ratio, and a first conversion processing means. After the original signal having the signal component of (1) is mirror-folded in the frequency direction at the Nyquist frequency before resampling, the DCT
Signal generating means for generating a second signal component shifted in the high frequency direction by one component of a coefficient; mirror image inverting means for mirror-inverting the second signal component so that the positive and negative of the frequency response are inverted; Filter means for performing a filtering process using a band-limiting filter until the Nyquist frequency later, wherein the gain adjusting means, the signal generating means, the mirror image inverting means, and the pseudo high frequency component generated by the filter means are shifted to a high frequency side. Is used as the DCT coefficient of According to another feature of the present invention, the signal generation means converts the second signal component into:
If the enlargement ratio is smaller than two times, the component up to the Nyquist frequency after the resampling is used. If the enlargement ratio is larger than two times, only one component of the DCT coefficient is returned after mirror image folding in the frequency direction. A third signal component generated by mirror-reflecting the component shifted in the high-frequency direction to the higher frequency side is added to make the component up to the Nyquist frequency after the resampling. Another feature of the present invention is that the band-limiting filter takes the frequency response on the vertical axis on the y-axis, the frequency on the horizontal axis on the x-axis, the response at the origin on the frequency axis as A, Nyquist frequency is fn
Is given by the following equation, y = AA * (x / fn) ^1.8 . Another feature of the present invention is that the band limiting filter is performed using a window function including a panning window and a Blackman window. Another feature of the present invention is that the second conversion processing means adjusts a gain according to the enlargement ratio, and then adjusts the gain to 0 on the high frequency side of the first signal component.
It is characterized in that a value is inserted. Another feature of the present invention is that the edge detecting means divides each DCT coefficient block into four subbands divided into a high frequency component and a low frequency component in the horizontal and vertical directions, respectively. And 4 divided by the dividing means.
And a component determining means for determining whether or not a DCT coefficient of a sub-band other than a low-band sub-band in the horizontal and vertical sub-bands includes a component larger than a preset threshold. It is characterized by:

The image enlarging apparatus according to the present invention comprises: a DCT means for DCT of an input image signal in a time-space domain;
An edge detection unit that detects edges in block units using a DCT coefficient that is DCT-converted by the CT unit; Processing means; inverse DCT means for performing an inverse DCT on the DCT coefficient block after the enlargement processing by the image processing means with a base matrix corresponding to the enlarged block size; The image processing apparatus further includes a system control unit that receives the edge information and controls the DCT unit, the image processing unit, and the inverse DCT unit based on information including an externally input image resolution and an output image resolution. Another feature of the present invention is that, of the two types of pixel number conversion processes performed by the image processing means, a pixel number conversion process is performed on a block detected to include the edge. Conversion processing means, and second conversion processing means for performing a pixel number conversion processing on a block detected not to include the edge, wherein the first conversion processing means Gain adjusting means for adjusting the gain, and mirror-folding the original signal having the first signal component in the frequency direction at the Nyquist frequency before resampling, and then shifting the DCT coefficient by one component in the high frequency direction. Signal generating means for generating a second signal component, mirror image inverting means for mirror-inverting the second signal component so that the frequency response is inverted, and Nike after resampling. Filter means for performing a filtering process using a band-limiting filter up to a frequency band, wherein the high-frequency side DCT is generated by the gain adjusting means, the signal generating means, the mirror image inverting means, and the filter means. It is characterized in that it is used as a coefficient. Another feature of the present invention is that the band-limiting filter takes the frequency response on the vertical axis on the y-axis, the frequency on the horizontal axis on the x-axis, the response at the origin on the frequency axis as A, The Nyquist frequency is represented by fn, and the following equation is given by y = AA * (x / fn) ^1.8 . Another feature of the present invention is that the band limiting filter is performed using a window function including a panning window and a Blackman window. Another feature of the present invention is that, of the two types of pixel number conversion processes performed by the image processing means, the pixel number conversion process is performed on a block that is detected as not including an edge, by using an enlargement ratio. After the gain is adjusted in accordance with, the zero value is inserted into the high frequency side of the original signal component. Another feature of the present invention is that the edge detecting means divides each DCT coefficient block into four subbands divided into a high frequency component and a low frequency component in the horizontal and vertical directions, respectively. And whether or not there is a component larger than a preset threshold value in the DCT coefficients of the subbands other than the subbands that are both low in the horizontal and vertical directions among the four subbands divided by the dividing means. And determining means for determining

The method for processing a television signal according to the present invention comprises:
After the first signal component, which is the frequency component of the input original signal, is mirrored in the frequency direction at the Nyquist frequency before resampling, the DCT coefficient of 1
A folding step of generating a second signal component shifted by a component in the high frequency direction, an inverting step of mirror-inverting the second signal component to generate a third signal generation amount, and a second inverting step generated by the inverting step. 3 is added as a high-frequency component of the input original signal to generate a pseudo high-frequency component, and the resampling is performed from the low-frequency component of the first signal component. And a filtering step of filtering the frequency components up to the Nyquist frequency band after the filtering. Another feature of the present invention is that a pseudo high frequency component up to the Nyquist frequency after sub-sampling the first signal component, which is a frequency component of the input original signal, is converted into a high frequency component of the original signal. And a filter step of filtering frequency components from the low-frequency component of the first signal component to the pseudo high-frequency component inserted in the insertion step. . According to another feature of the present invention, a variable length decoding step of performing variable length decoding on an encoded bit stream, and a quantization DCT of a predetermined block decoded by the variable length decoding step.
An inverse quantization step of inversely quantizing the coefficients; an edge detection step of detecting edges in block units using the quantized DCT coefficients restored in the inverse quantization step; and an edge detection step in which the edge detection step includes an edge. An image processing step of performing different enlargement processing on the first block that has been detected and the second block that has been detected as not including an edge; and enlarging the DCT coefficient block that has been image-processed by the image processing step. An inverse DCT step of performing an inverse DCT with a base matrix corresponding to the set block size, and performing a local decoding by motion compensation in consideration of the conversion of the number of pixels to obtain a video signal before the bit stream is encoded. The decoding step of decoding and the number of pixels of the video signal processed by the above-mentioned decoding step become an appropriate number of pixels to be displayed on the display means. A decoder having a number-of-pixels converting step of converting the video signal into a video signal; and a system for controlling the content of processing performed in the decoder step based on information including image resolution and edge information of the video signal processed by the decoder step. And a display control step of converting a signal processed by the decoder step into a video signal displayable in the display step and controlling a display process in the display step. According to another feature of the present invention, the image processing step includes a first conversion step of performing a pixel number conversion process on a block detected to include the edge, and an image processing step that does not include the edge. A second conversion step of performing a pixel number conversion process on the detected block,
The first conversion step includes a gain adjustment step of adjusting a gain according to an enlargement ratio, and mirror-folding of an original signal having a first signal component in a frequency direction with a Nyquist frequency before resampling as a boundary. Thereafter, a signal generation step of generating a second signal component shifted in the high frequency direction by one component of the DCT coefficient, and a mirror image inversion step of mirror-inverting the second signal component so that the sign of the frequency response is inverted. A filter step of performing a filter process using a band-limiting filter up to the Nyquist frequency after the resampling, and the pseudo high band component processed by the gain adjustment step, the signal generation step, the mirror image inversion step, and the filter step. It is characterized in that it is used as a DCT coefficient on the high frequency side. Another feature of the present invention is that, in the signal generation step, the second signal component is a component up to the Nyquist frequency after the resampling when the enlargement ratio is smaller than 2, If the magnification ratio is larger than twice, after mirror image folding in the frequency direction, D
A third signal component generated by mirror-reflecting a component shifted in the high frequency direction by one component of the CT coefficient to the higher frequency side is added to make a component up to the Nyquist frequency after the resampling. . Another characteristic of the present invention is that the characteristics of the band-limiting filter are such that the frequency response on the vertical axis is on the y-axis, the frequency on the horizontal axis is on the x-axis, and the response at the origin of the frequency axis is A, It is characterized by the following equation, y = AA * (x / fn) ^1.8, where fn is the Nyquist frequency after resampling. Another feature of the present invention is that the band-limiting filter is a panning window,
It is characterized by being performed using a window function including a Blackman window. Another feature of the present invention is that, in the second conversion step, after adjusting a gain according to the enlargement ratio, a 0 value is inserted into a high frequency side of the first signal component. It is characterized by the fact that: According to another feature of the present invention, the edge detecting step includes:
A dividing step of dividing each DCT coefficient block into four sub-bands divided into a high-frequency component and a low-frequency component in the horizontal and vertical directions, respectively, of the four sub-bands divided by the dividing step, And a component determining step of determining whether or not a DCT coefficient of a sub-band other than the low-band sub-band has a component larger than a preset threshold.

The image enlarging method according to the present invention includes a DCT step for DCT of an input image signal in a time-space domain;
An edge detection step of detecting edges in block units using the DCT coefficients DCT obtained by the CT step, an image subjected to different enlargement processing for a block detected to include an edge and a block not including an edge in the edge detection step Processing step, an inverse DCT step of performing an inverse DCT on the DCT coefficient block that has been enlarged by the image processing step with a base matrix corresponding to the enlarged block size, and an inverse DCT step for each block from the edge detection step. And a system control step of receiving the edge information and controlling the DCT step, the image processing step, and the inverse DCT step based on information including an externally input image resolution and an output image resolution. Another feature of the present invention is that, of the two types of pixel number conversion processes performed in the image processing step, the first pixel number conversion process is performed on a block detected to include the edge. And a second conversion step of performing a pixel number conversion process on the block detected as not including the edge, wherein the first conversion step adjusts a gain according to an enlargement ratio. A signal adjusting step of performing mirror image folding of an original signal having a first signal component in a frequency direction with a Nyquist frequency before resampling as a boundary to generate a second signal component; A mirror image inverting step of inverting the signal component of the signal so that the positive and negative of the frequency response are inverted, and a filter for performing a filtering process using a band-limiting filter up to the Nyquist frequency after the resampling. And a data process is characterized by using the gain adjustment step, the signal generation step, mirror-reversing step, and a pseudo high frequency components generated by the filter process as DCT coefficients of the high frequency side. According to another feature of the present invention, the band limiting filter processing is such that the frequency response on the vertical axis is on the y-axis, the frequency on the horizontal axis is on the x-axis, the response at the origin of the frequency axis is A, The following equation, y = AA * (x
/ fn) It is characterized by processing based on the characteristics given in ^1.8 . Another feature of the present invention is that the band limiting filter processing is performed using a window function including a panning window and a Blackman window. Other features of the present invention include:
Among the two types of pixel number conversion processing performed in the image processing step, the pixel number conversion processing is performed on a block that is detected as not including an edge, after adjusting the gain according to the enlargement ratio, and adjusting the high of the original signal component. It is characterized in that it is performed by inserting a 0 value on the band side. According to another feature of the present invention, the edge detecting step includes the step of:
A dividing step of dividing the coefficient block into four sub-bands divided into a high-frequency component and a low-frequency component in the horizontal and vertical directions, respectively, of the four sub-bands divided by the dividing step, A determining step of determining whether a DCT coefficient of a sub-band other than the low-band sub-band includes a component larger than a predetermined threshold value.

A storage medium according to the present invention is characterized in that a program constituting each means described above is stored so as to be readable from a computer. Another feature of the present invention is that a program for executing the above-described method is stored in a computer-readable manner.

[0029]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS (Purpose of the Invention) First, DCT coefficient blocks containing a large amount of high frequency components in spatial frequencies are performed. In the enlargement process of the present embodiment, the purpose of reducing ringing generated near the edge is as follows.
This will be described with reference to FIGS.

FIG. 5 is a conceptual diagram for explaining a case where the input signal is enlarged twice. It is assumed that the spatial frequency characteristic of the input signal in FIG. 5A includes a large amount of high-frequency components as shown by the curve a, that is, that many responses exist near the Nyquist frequency.

In the conventional example, a high band from the Nyquist frequency fn in FIG. 5A to the Nyquist frequency fn ′ after sub-sampling in FIG. 5B (the position of the sampling frequency fs in FIG. 5A) is set. Since “0 value” has been inserted, there is no high-frequency component after sub-sampling. For this reason, ringing due to the lack of the high-frequency component has occurred near the edge region of the image returned by the IDCT (inverse cosine transform) in the time-space region.

In the present embodiment, the original signal of the curve a is represented by fn '
/ 2 (i.e., fn), the frequency response is shifted in the frequency direction by one component of the DCT coefficient (curve a 'in FIG. 5A), and the sign of the frequency response is further inverted. Folded.

By adding the curve b in FIG. 5B as a high-frequency component, a pseudo high-frequency component is generated, and a band-limiting filter process is performed as shown by the curve in FIG. 5C. As a result, it is possible to reduce the ringing near the edge due to the lack of the high frequency component. Note that a curve b ′ in FIG. 5C shows the curve b after the band limiting filter processing.

FIG. 6 is a conceptual diagram for explaining a case where the enlargement ratio is larger than 1 and smaller than 2 times, for example, 3/2 times. In this case, the pseudo high frequency component is additionally inserted on the high frequency side up to the Nyquist frequency fn 'after the sub-sampling as shown by a curve b in FIG. An example is shown in which band limiting filter processing as shown by the curve in FIG. 6C is performed.

FIG. 10 shows an image after the enlargement processing in the case of 3/2 times. 10A shows an original image, FIG. 10B shows a conventional DCT method, and FIG. 10C shows an image enlarged by the present method.

FIG. 7 shows that when the enlargement ratio exceeds 2 times,
It is a conceptual diagram explaining about the case of 5/2 times, for example.
In this case, as in the case of the double, the pseudo high frequency component is additionally inserted up to the band of fs in FIG. 7A, and the higher frequency band up to fn ′ in FIG. 7B is further increased. 7A, a component obtained by folding the pseudo high frequency component in the frequency direction at fs in FIG. 7A is additionally inserted.
Curve b) in (b).

FIG. 7C shows the curve. The band limiting filter processing shown in FIG. FIG. 11 shows an image after the enlargement processing in the case of 5/2 times. FIG. 11A shows an original image, FIG. 11B shows a conventional DCT method, and FIG. 11C shows an image enlarged by the present method. From this image, it can be easily confirmed that ringing near the edge, which has been a problem in the conventional method, has been reduced.

Similarly, it goes without saying that the present invention can be applied to a case where the enlargement ratio exceeds three times. When the frequency response on the vertical axis is set on the y-axis and the frequency on the horizontal axis is set on the X-axis, and the response at the origin of the frequency axis is A, for example, the following equation (1) shows a curve of γ = 1.8. Given. y = A−A * (x / fn ′) ^1.8 (1) The value of γ = 1.8 has been obtained as a suitable value by experiment, and may be another value.

Furthermore, a Hamming window, Bl
Even if a window function such as the ackman window is used, no significant deterioration is observed, but it is undeniable that the attenuation of higher frequency components becomes stronger.

<First Embodiment> A first embodiment of the present invention will be described below with reference to the drawings. However, hereinafter, a solid line represents a data line, and a broken line represents a control line.

FIG. 2 is a block diagram showing a configuration example of a television receiver according to the first embodiment of the present invention. This block comprises a decoder 21, a display controller 22, a system controller 23, and a display 24.

A transport stream input from a tuner means (not shown) or a storage type digital device (not shown) is decoded by a decoder means 21 and is controlled by a system control means 23. , The number of pixels is converted as described above.

The output signal from the decoder 21 is converted by the display controller 22 into a signal that can be displayed on the display 24.

<Decoder Means> First, the decoder means 21 according to the first example of this embodiment is shown in FIG.
This will be described with reference to the block diagram of FIG.

The above block comprises a buffer means 11, a variable length decoding means 12, an inverse quantization means 13, and an image processing means 1.
4. It comprises an edge detecting means 15, an inverse DCT means 16, a motion compensation predicting means 17, a video memory means 18, and a format converting means 19.

First, the structure of the input coded data, that is, the bit stream of MPEG is also shown in FIG.
This will be described with reference to FIG. As shown in FIGS. 8 and 9, the MPEG bit stream includes a sequence layer including a sequence header 81, a sequence extension 82, an extension and user data (0) 83, and a GOP (Group).
of picture) header 84, a GOP layer including extension and user data (1) 85, and picture header 8
6, picture coding extension 87, extension and user data (2) 88, picture layer consisting of picture data 89, sequence end 810, sequence header 81
1, picture data composed of slice data 812,
It is composed of a slice layer composed of slice information 813 and macroblock data 814, and a macroblock layer composed of macroblock information 815 and block data 816.

(Sequence Layer) The sequence layer starts with a sequence header, and an extension of the sequence extension 82.
The bit stream of MPEG-1 and MPEG-2 is classified according to the presence or absence of the start code.

In general, a sequence represents an entire video program and ends with sequence encoded. The sequence layer basically includes one or more GOPs. The sequence header includes information set for each sequence, such as an encoded image size, an aspect ratio, a frame rate, a bit rate, a VBV buffer size, and a quantization matrix.

(GOP Layer) The GOP layer starts with a GOP header and is composed of one or more pictures. The first coded video of the GOP layer is independently coded without using a reference screen. Become a picture.

Therefore, by using an I picture, a GOP can be used as a point for performing random access from MPEG data. Note that applications such as communication require low delay characteristics, and therefore, in MPEG-2, the GOP layer can be omitted.

In this case, for example, an intra-slice to be intra-coded in slice units, which will be described later, is used, and the intra-coding of each slice is circulated by dividing into several screens. The occupancy can also be reduced on average.

(Picture Layer) The picture layer corresponds to each screen, and each picture layer is classified into one or more slice layers. In the picture header 86, encoding conditions for the screen are set. In the picture coding extension 87, the motion vector ranges in the front-back direction and the horizontal and vertical directions are specified, and the frame structure and the field structure are set.

Further, the setting of the DC coefficient accuracy of the intra macroblock, the selection of the VLC (variable length code) type, the selection of the linear or non-linear quantization scale, the selection of zigzag, and the alternate scanning are performed.

(Slice Layer) The slice layer indicates a horizontally long band-like area in the screen. By forming the screen with a plurality of slices, even if an error occurs in a certain slice layer,
Error recovery is possible by synchronization from the start of the next slice layer.

The slice layer is composed of one or more macroblocks, arranged in a raster scan order from left to right and from top to bottom. The length and start position are free and can be changed for each screen. However, in MPEG-2, one slice does not extend over a downward direction.

(Macro Block Layer) The macro block is
For example, in the case of the “4: 2: 0” format, the format includes six blocks of four luminance blocks and two chrominance blocks. In the macroblock data, the position of the macroblock and the encoding mode are set.

(Block layer) A block is composed of 8 pixels × 8 lines of a luminance signal or a color difference signal, and DCT (discrete cosine transform) and IDCT (inverse discrete cosine transform) are performed in this unit. Block data is quantized D
It is composed of CT coefficients.

For the DC component of the intra macroblock, the magnitude and difference information relating to the difference value with respect to the adjacent block are given, and for the other DCT coefficients, information relating to the length and level of 0 coefficient up to the non-zero quantized DCT coefficient. Is given, and the DCT coefficient of each block ends at EOB (End of Block).

Next, the processing of each block in FIG. 1 will be described. (Variable Length Decoding Unit) The variable length decoding unit 12 reads the coded data buffered by the buffer unit 11, decodes the macroblock coding information, and sets the coding mode, motion vector, quantization information, Generalized DCT coefficients are separated. Note that the variable-length coding performed on the encoder side is performed by assigning a shorter code to data having a higher appearance frequency, and the reverse process is performed in the variable-length decoding.

(Inverse quantization means) In the inverse quantization means 13,
The decoded 8 × 8 quantized DCT coefficients are inversely quantized to obtain D
Restore to CT coefficients. In the quantization, the encoder compresses spatial information using a quantization table determined according to the visual characteristics of a person, and the inverse quantization performs the inverse process by inverse quantization. This is performed using a conversion table.

(Image processing means) In the image processing means 14,
Under the control of the system control means 23 described later, the number of horizontal and vertical pixels of the input signal is set to be equal to the number of horizontal and vertical pixels of the display means 24 described later, or the desired number of horizontal and vertical pixels as described above. DCT coefficient conversion is performed so that

At this time, gain adjustment according to the enlargement ratio is performed on the components of the input signal. For example, if the horizontal and vertical enlargement ratios are both double, the gain of the input DCT coefficient is corrected to double.

In the present embodiment, a block including an edge and a block not including the edge are determined by an edge detection unit 15 described later, and different processing is performed according to the determination result.

<Case of Block Determined to Include Edge> As described in the gist of the present embodiment, a frequency component higher than the frequency component of the original signal is pseudo-generated based on the original signal. Then, the band is limited according to the block size after the enlargement. Detailed processing has been described in detail in the gist of the present embodiment, and a description thereof will not be repeated.

<In the case of a block determined not to include an edge> In this case, processing is performed by the conventional method, that is, further outside the right side and lower side of the frequency component of the original signal (FIG. 4 (b) 0). By additionally inserting “0 value” into the (gray part), an enlarged DCT coefficient block as shown in FIG. 4 is output to the inverse DCT means 16 described later.

The expanded DCT coefficient is subjected to inverse DCT transform by a base matrix having a size equal to the enlarged DCT coefficient size in an inverse DCT means 16 to be described later.
As a result, the number of pixels in the spatial domain is converted.

(Edge Detection Means) Edge block detection performed by the edge detection means 15 will be described with reference to the flowchart of FIG. 14 and the conceptual diagram of FIG. The DCT coefficient block output from the inverse quantization means 13 (for example, the luminance signal is 16 × 16, the color difference signal is 8 ×
8) is divided into four frequency bands (sub-bands) of a high frequency component and a low frequency component in each of the horizontal and vertical directions (S)
131).

FIG. 17A shows that the high-frequency and low-frequency components are “1: 1”.
FIG. 17B shows an example in which the high-frequency and low-frequency components are divided into “3: 1” (the gray area in the figure is the high-frequency component). Of these, DCT coefficients of three areas other than the low-frequency area in both the horizontal and vertical directions (gray area in FIG. 17)
It is determined whether or not there is a component larger than the predetermined threshold value (S132).

If it is determined that the block exists, the block is determined to be a block including an edge (S13).
3). If no block exists, it is determined that the block does not include an edge (S134). Finally, it is determined whether or not the processing has been completed for all the blocks (S135). If the processing has not been completed, the process returns to S131 to repeat the above-described processing. The high-frequency component and the low-frequency component may be separated from each other by half of the DCT coefficient block size of the original signal, or may be other than that.

(Inverse DCT Means) The inverse DCT means 16 receives the control of the system control means 23, which will be described later, and converts the DCT coefficient blocks whose block sizes have been converted into DCT base matrices corresponding to the sizes of the respective blocks. Inverse DCT transform to convert the DCT area into pixel space area data.

For example, offset data and rounding processing are performed on 8-bit precision integer value data. D as described above
Although the CT and the inverse DCT operation are defined as real number operations, the result of the inverse DCT operation does not match the DCT input value and does not always become an integer because quantization processing is performed between them.

For this reason, when the value after the decimal point of the calculation result is "0.5", there are cases where the value is rounded up and down when converting to an integer. Since this mismatch cannot be solved even if the operation precision is specified as follows,
As a countermeasure, the coefficient value after the inverse quantization is minutely changed to reduce the probability that the value after the decimal point of the inverse DCT operation result without error is “0.5”.

(Motion Compensation Prediction) Next, the motion compensation prediction means 17 will be described in detail with reference to FIG. FIG.
, 125 is a predicted field / frame selecting means, 126 is a frame store address designating means, 127
Denotes a frame memory unit, 128 denotes a half-pixel prediction filtering unit, 129 denotes a prediction combination unit, 1210 denotes an addition unit, and 1211 denotes a saturation processing unit.

In the case of the motion compensation prediction mode, the motion compensation prediction means 17 outputs the block data (p
[Y] [x]) is output from the inverse DCT means 16 (f
[Y] [x]).

However, in the case of the intra coding mode, this processing is not performed. Therefore, (p [y]
[X]) is zero. Note that this coding mode is determined for each macroblock unit, the motion compensation prediction mode can be expected to have high coding efficiency when the temporal correlation is high, and the intra coding mode is for large changes in scenes, etc.
Used when temporal correlation cannot be expected.

Also, if a block is not coded, there is no coefficient data either because the entire macroblock has been skipped or a particular block has not been decoded.

In this case, (f [y] [x]) is zero, and the decoded pixel is simply predicted (p [y]
[X]). Further, a saturation processing unit 1211 is required so that (f [y] [x]) does not become zero.

First, the motion compensation prediction mode will be described. The prediction modes are field prediction and frame prediction.
Broadly classified into types. In the field prediction, data from one or a plurality of previously decoded fields is used, and prediction is independently performed for each field. In the frame prediction, a frame is predicted from “1” or a plurality of frames decoded earlier. It is necessary to understand that the fields and frames on which the prediction is based are themselves decoded as either field images or frame images.

In a field image, all predictions are field predictions. On the other hand, in a frame image, either field prediction or frame prediction can be used (selected for each macroblock). In addition to the field or frame prediction, two special prediction modes of 16 × 8 motion compensation and dual prime are used.
In 16 × 8 motion compensation, two sets of motion vectors are used for each macroblock.

The first vector is at the top of the 16 × 8 region,
The second vector is used below the 16x8 area. In the case of a macroblock predicted in both directions, two sets of vectors are used for forward prediction and two sets of vectors are used for backward prediction, for a total of four vectors.

In the dual prime compensation system, one vector is encoded together with a small difference vector in a bit stream. In the case of a field image, two sets of motion vectors can be generated from this information.

These vectors are used to form predictions from two reference fields (one at the top, one at the bottom) and are averaged to form the final prediction. In the case of a frame image, this process is repeated for two fields so that prediction from a total of four fields is performed.

This prediction mode can be used only for a P picture, and there is no B picture between a picture to be predicted and a reference field or frame. Predicted field / frame selection unit 125 makes a selection of whether to use which field and frame for prediction formation.

The field prediction in the P picture is performed using the reference top field and the reference bottom field which are encoded immediately before. Field prediction in a B picture is performed from two fields of two reference frames restored immediately before.

On the other hand, the frame prediction in the P picture is performed from the reference frame restored immediately before. Similarly B
The frame prediction in the picture is the previously restored 2
From one reference frame.

The color component scaling means 124 obtains the motion vector of the color difference signal by scaling the motion vector of the luminance signal. For example, in the case of the 4: 2: 0 format, the motion vector is halved in both the horizontal and vertical directions.

The half-pixel prediction filtering means 128
Double the motion vector. This is because, in order to improve the prediction accuracy, 0.5 pixel accuracy is used in which each pixel on the reference screen is linearly interpolated 1: 1.

Next, the motion vector coding will be described. Utilizing the characteristic that the motion amount in the vicinity of the image has a high correlation, the coding of the motion vector
The motion vector amount of the previously encoded macroblock is used as the prediction vector PMV, and the difference vector delta from the prediction vector PMV is encoded.

Motion vector vector '[r] [s] [t]
Is expressed as Expression (2), and vector ′ [r] [s] [t] = PMV [r] [s] [t] + delta (2) Information about the difference vector delta is encoded.

Here, r is the first / second macro block.
, S indicates a forward / backward motion vector, and t indicates a horizontal / vertical component. Further, the difference vector delta is represented by scaling the basic difference vector motion code by the scale factor f and adding the residual vector motion residual, as shown in Expression (3). Delta = Sign (motion code) * [((Abs (motion code) −1) * f) + motion residual + 1｝ (3)

However, Sign indicates a sign, and Abs indicates an absolute value. Here, the basic difference vector motion code is “−
An integer from "16" to "+16", which is given to the macroblock as a variable length coded code (1 to 11 bits).

The scale factor f is a scale factor code f # code [s] for determining a motion compensation range.
Using [t], f = 1 << (f # code [s] [t] -1) (4) Here, << indicates a left bit shift. The residual vector motionresidual is given to the macroblock and has a bit length of (fcode-1).

Therefore, what is encoded in the macro block as the motion vector information is the basic difference vector mo.
The motion code and residual vector motion residual, for example, scale factor fcode = 8, basic difference vector mot
When the ion code = −16 and the residual vector motion residual are given as 127, the difference vector delta = −2048 using the above equations (3) and (4).

As described above, the difference vector delta has a scale factor code and a motion vector having 33 motion vectors.
A wide range can be efficiently coded using the code table and the residual vector.

By the way, in MP @ ML, in the case of frame motion compensation prediction of a frame structure, the range of the motion vector is [−128.0, +127.5] in the vertical direction and [−1024.0, 1023. 5] (fcod is 8 and 5, respectively).

In this embodiment, the inverse DCT
Since the output from the means 16 has been subjected to enlargement processing (conversion of the number of pixels) by the image processing means 14 described above, the output from the means 16 is controlled by the system control means 23 which will be described later, and the number of horizontal and vertical pixels after scaling is considered. Perform motion compensation on above.

That is, for example, when scaling is performed twice, the size of the macroblock is doubled from 16 × 16 (in the case of a luminance signal) to 32 × 32 in both the horizontal and vertical directions. Since the size of one picture and that of a P picture are also doubled, the amount of the motion vector vector '[r] [s] [t] must naturally be doubled both horizontally and vertically.

(Video memory) Video memory means 18
Saves an I picture and a P picture as reference pixels used in the decoding process.

(Format Conversion) Format Conversion 1
Reference numeral 9 denotes an I (Intra coded) picture, P (Predictive) rearranged on the decoder side to increase the encoding efficiency.
coded) picture (b) (Bidirectionally Predictive
coded) picture in the original input order.
Then, the image size is converted and output as necessary.

In this embodiment, since the output from the inverse DCT means 16 is scaled (the number of pixels is converted), the format conversion 19 is also controlled by the system control means 23 which will be described later, and The rearrangement is performed in consideration of the number of vertical pixels.

<System Control Means> Next, the system control means 23 will be described with reference to the flowchart of FIG. The system control means 23 includes the decoder means 2
The number of horizontal and vertical pixels (hereinafter, referred to as in # X and in # y) of the input signal is obtained from the sequence header and the sequence extension of the coded bit stream input to 1 (S141).

However, in the horizontal and vertical pixel numbers, the lower 12 bits in both the horizontal and vertical pixels are included in the sequence header and the upper 2 bits.
Bits are in the sequence extension. For reference, FIG.
20 and FIG. 20 show the contents of the sequence header and the sequence extension, respectively.

Then, the number of horizontal and vertical pixels (hereinafter referred to as Out # X and Out # y) of the display means 24 described later is divided by the number of horizontal and vertical pixels in # X and in # y of the input signal, respectively. Horizontal and vertical scaling (pixel number conversion) ratio (hereinafter, sr # x,
sr # y) and an output block size M × N obtained by integrating each of the above sr # x and sr # y into an input block size 8.
(M and N are integers equal to or greater than 8) are calculated (S142).

Then, a setting is made so that the gain adjustment according to the sr # x and sr # y is performed on the input DCT coefficient (S143). Here, the edge detecting means 15 described above is used.
In the case of a block determined not to include an edge (S144) in step S144, the image processing means 14 sets "0" in the lower right and right sides (high frequency components) of the DCT coefficients of the original signal.
Value ”is added and the block size after addition is M × N
(S145).

On the other hand, in the case of a block determined to contain an edge by the edge detecting means 15, the image processing means 14 applies the lower right and right sides (high frequency components) of the DCT coefficients of the original signal of the present embodiment to the above-described embodiment. The pseudo high frequency component described above is additionally inserted, and the block size after the addition is set to be M × N (S146).

Further, the image processing means 14 is set so that the band limiting process is performed in the horizontal and vertical directions in accordance with the block size of M × N as described in the purpose of the present embodiment (S147).

Further, the basis matrix size of the inverse DCT transform performed by the inverse DCT means 16 is M ×
N is set (S148). Then, the motion compensation processing in the motion compensation prediction means 17 described above is performed by scaling the number of horizontal and vertical pixels to Out # X and Out # y, respectively, and the amount of motion vector to vector [r] [s] [0]. #x, vector
[R] [s] [1] is set to use a value scaled by sr # y (S149).

Further, the setting is made such that the above-described frame or field rearrangement processing in the format conversion means 1.9 is performed with the values of the horizontal and vertical pixels being Out # X and Out # y, respectively (S1410).

The above is the control of each means in the decoder means 21. Further, the display control means 22 is also set so that the number of horizontal and vertical pixels becomes Out # X and Out # y, respectively ( S1411).

<Second Embodiment> Hereinafter, a second embodiment of the present invention will be described with reference to the drawings. In the present embodiment, an input signal encoded by MPEG2 is received,
In this embodiment, the moving image is enlarged by performing signal processing on the basis of the DCT coefficient after the inverse amount pre-compensation in the decoder means 21 for decoding this.

In the second embodiment, after a digital signal input in a more general configuration, that is, an uncoded ordinary time-space domain is converted into a DCT domain, the above-described first embodiment is performed. The enlargement processing is realized in the same manner as in the above.

FIG. 16 is a block diagram showing a configuration example of an image enlargement apparatus according to the second embodiment. The image enlargement apparatus includes a DCT unit 151, an edge detection unit 152, a system control unit 153, an image processing unit 154, an inverse DCT
Means 155.

(DCT means) The DCT means 151 receives control from the system control means 153 described later,
DCT is performed on the image input signal in the spatial domain using a desired basis matrix, and DCT coefficients are output in block units.

(Edge detection means) Edge detection means 152
Is based on the DCT coefficient output from the DCT means,
It is determined whether or not an edge (high-frequency component) is included in block units, and the result is sent to a system control unit 153 described later.

The details of the processing performed by the edge detecting means are the same as those in the first embodiment, and a detailed description thereof will be omitted.

(Image processing means) The image processing means 154 converts DCT coefficients so that the input resolution is converted to a desired output resolution under the control of the system control means 153 described later. Also in the present embodiment, the edge detection unit 152 determines a block including an edge and a block not including the edge, and performs different processing according to the determination result.

The details of the processing performed by the image processing means 154 are the same as in the first embodiment, and will not be described.

(Inverse DCT Means) The inverse DCT means 155
Under the control of the system control unit 153 described later, the DCT coefficient block whose block size has been converted by the image processing unit 154 is subjected to inverse DCT using a DCT base matrix corresponding to each converted block size. The area is converted into pixel space area data and output.

<System Control Means> Next, the system control means 153 will be described with reference to the flowchart of FIG. The system control unit 153 determines the input / output block size from the input and output resolutions input from outside the present image enlargement device.

However, in the case of a normal YCbCr input, that is, a luminance signal and a color difference signal input, the input block size of the luminance signal is 16 × 16 and the input block size of the color difference signal is 8 × 8.

The output block size calculated by multiplying the input block size by the horizontal and vertical enlargement ratios is M ×
N (S171). Then, it is set so that the gain adjustment according to the horizontal / vertical enlargement ratio is performed on the input DCT coefficient (S172).

Next, edge information for each block is obtained from the edge detection means 152 (S173). If the block is determined not to include an edge, the image processing means 154 determines the DCT coefficient of the original signal. “0 value” is additionally inserted in the lower right and right sides (high frequency components), and the block size after the addition is set to be M × N (S17).
5).

On the other hand, in the case of a block determined to contain an edge by the edge detecting means 155, the image processing means 154 applies the lower right and right sides (high frequency components) of the DCT coefficients of the original signal of the present embodiment to the above. The pseudo high-frequency component described for the purpose is additionally inserted, and the block size after the addition is set to be M × N (S176).

Then, the image processing means 154 is set so that the band limiting process is performed in the horizontal and vertical directions according to the block size of M × N as described in the purpose of the present embodiment (S177). .

Further, the base matrix size of the inverse DCT transform performed by the inverse DCT means 155 is M
× N is set (S178).

(Other Embodiments of the Present Embodiment) The present embodiment is applied to a system constituted by a plurality of devices (for example, a host computer, an interface device, a reader, a printer, etc.) and is constituted by one device. You may apply to an apparatus.

Further, the functions of the above-described embodiment are implemented in an apparatus connected to the above-described various devices or a computer in a system so that various devices are operated so as to realize the functions of the above-described embodiment. Supplies the software program code for performing the
The present invention also includes those implemented by operating the various devices according to the program stored in U).

Further, in this case, the program code of the software implements the functions of the above-described embodiment, and the program code itself and means for supplying the program code to the computer, for example, the related program The storage medium storing the code constitutes this embodiment. As a storage medium for storing the relevant program code, for example, a floppy (registered trademark) disk, hard disk, optical disk, magneto-optical disk CD-ROM, magnetic tape, nonvolatile memory card, ROM, or the like can be used.

When the computer executes the supplied program code, not only the functions described in the above-described embodiment are realized, but also the OS (operating system) or the operating system running on the computer. It goes without saying that the program code related to the case where the functions described in the above-described embodiments are realized in cooperation with other application software or the like is included in the embodiments of the present embodiment.

Further, after the supplied program code is stored in the memory provided in the function expansion board of the computer or the function expansion unit connected to the computer, the function expansion board or the function expansion unit is stored based on the instruction of the program code. This embodiment is also included in the case where a CPU or the like provided in the device performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0131]

As described above, according to the present invention, according to the present invention, "0 value" is inserted into a conventional high-frequency component, so that high-definition video contents containing many edges can be provided with high quality. It is possible to provide an expandable television receiver and an image enlargement device.

[Brief description of the drawings]

FIG. 1 is a block diagram for explaining a decoder unit according to a first embodiment of the present invention.

FIG. 2 is a block diagram for explaining an example of the overall configuration according to the first embodiment of the present invention.

FIG. 3 is a block diagram for explaining conventional decoder means.

FIG. 4 is a conceptual diagram for explaining an enlargement process in a DCT area according to a conventional example.

FIG. 5 is a conceptual diagram for describing generation of a pseudo high-frequency component according to the embodiment of the present invention.

FIG. 6 is a conceptual diagram for describing generation of a pseudo high frequency component according to the embodiment of the present invention.

FIG. 7 is a conceptual diagram for explaining generation of a pseudo high frequency component according to the embodiment of the present invention.

FIG. 8 is a block diagram for explaining bit stream data input to a decoder unit according to the first embodiment.

FIG. 9 is a block diagram for explaining bit stream data input to a decoder unit according to the first embodiment.

FIG. 10 is a conceptual diagram illustrating a result of an image enlargement process according to the embodiment.

FIG. 11 is a conceptual diagram showing a result of an image enlargement process according to the embodiment.

FIG. 12 is a conceptual diagram illustrating a conversion example of the number of pixels according to the embodiment of the present invention.

FIG. 13 is a block diagram for explaining a ringing motion compensation process according to the first embodiment.

FIG. 14 is a conceptual diagram for describing edge detection processing according to the embodiment of the present invention.

FIG. 15 is a flowchart for describing a system control unit according to the first embodiment.

FIG. 16 is a block diagram illustrating a configuration example of an image enlargement device according to a second embodiment.

FIG. 17 is a conceptual diagram illustrating edge detection processing according to the second embodiment.

FIG. 18 is a flowchart illustrating a system control unit according to the second embodiment.

FIG. 19 is a diagram for describing a sequence header in an MPEG2 bit stream.

FIG. 20 is a diagram for describing sequence extension in an MPEG2 bit stream.

[Explanation of symbols]

 Reference Signs List 11 buffer means 12 variable length decoding means 13 inverse quantization means 14 image processing means 15 edge detection means 16 inverse DCT means 17 motion compensation prediction means 18 video memory means 19 format conversion means 21 decoder means 22 display control means 23 system control means 24 display means 31 buffer means 32 variable length decoding means 33 inverse quantization means 34 inverse DCT means 35 motion compensation prediction means 36 video memory means 37 format conversion means 81 sequence header 82 sequence extension 83 extension and user data (0) 84 GOP Header 85 Extension and user data (1) 86 Picture header 87 Picture coding extension 88 Extension and user data 89 Picture data 810 Sequence end 811 Sequence header 812 Statuser 813 Slice information 814 Macroblock F data 815 Macroblock information 816 Block data 121 Vector prediction value 122 Vector encoding 123 Additional dual prime operation 124 Color component scaling 125 Prediction field / frame selection means 126 Frame store address designating means 127 Frame memory 128 Half pixel prediction filtering means 129 Prediction combination means 1210 Addition means 1211 Saturation processing means 151 DCT means 152 Edge detection means 153 System control means 154 Image processing means 155 Inverse DCT means

Claims (32)

[Claims] 1. After the first signal component, which is the frequency component of the input original signal, is mirror-reflected in the frequency direction from the Nyquist frequency before resampling,
Folding and shifting means for generating a second signal component shifted in the high-frequency direction by one component of the DCT coefficient; inverting means for mirror-inverting the second signal component to generate a third signal generation; High-frequency component generation means for adding the third signal component generated by the means as a high-frequency component of the input original signal to generate a pseudo high-frequency component;
A television receiver comprising: filter means for filtering a frequency component from a low-frequency component of the first signal component to a Nyquist frequency band after the resampling. 2. Insertion means for additionally inserting a pseudo high frequency component up to the Nyquist frequency after sub-sampling the first signal component, which is a frequency component of the input original signal, on the high frequency side of the original signal. 2. The television according to claim 1, further comprising: a filter unit configured to filter a frequency component from a low band component of the first signal component to a pseudo high band component inserted by the insertion unit. John receiver. 3. A variable length decoding means for performing variable length decoding on an encoded bit stream, and an inverse quantization means for inversely quantizing a quantized DCT coefficient of a predetermined block decoded by the variable length decoding means. Means, edge detecting means for detecting an edge in block units using the quantized DCT coefficients restored by the inverse quantizing means, and a first block detected to include an edge by the edge detecting means;
An image processing unit that performs different enlargement processing with the second block that is detected as not including an edge, and a DCT coefficient block that has been subjected to image processing by the image processing unit, in accordance with the enlarged block size. Inverse DCT means for performing inverse DCT on the basis matrix, and decoding means for performing local decoding by motion compensation in consideration of the conversion of the number of pixels and decoding the bit stream into a video signal before being encoded. A decoder having pixel number conversion means for converting the number of pixels of the video signal into an appropriate number of pixels for display on a display means; and an image resolution and an edge of the video signal output from the decoder. System control means for controlling the operation of the decoder means based on the information including information, and a signal output from the decoder means Converts into a displayable video signal to the serial display means, the television receiver characterized by having a display control means for controlling said display means. 4. The image processing means includes: first conversion processing means for performing a pixel number conversion process on a block detected to include the edge; and a first conversion processing means for performing a pixel number conversion process on a block detected not to include the edge. A second conversion processing unit for performing a pixel number conversion process, wherein the first conversion processing unit includes a gain adjustment unit for adjusting a gain according to an enlargement ratio, and an original signal having a first signal component. After performing mirror image folding in the frequency direction at the Nyquist frequency before resampling as a boundary, the DCT coefficient 1
Signal generating means for generating a second signal component shifted by a component in the high-frequency direction; mirror image inverting means for mirror-inverting the second signal component so that the positive or negative frequency response is inverted; and Nyquist after the resampling. Filter means for performing filter processing using a band-limiting filter up to a frequency, wherein a DCT coefficient of a high-frequency side of the pseudo high-frequency component generated by the gain adjustment means, the signal generation means, the mirror image inversion means, and the filter means is provided. The television receiver according to claim 3, wherein the television receiver is used as a television receiver. 5. The signal generating means according to claim 2, wherein said second signal component is a component up to the Nyquist frequency after said resampling when said enlargement ratio is smaller than 2, and said second signal component when said enlargement ratio is larger than 2. In the third example, a mirror image is folded back in the frequency direction, and a component shifted by one component of the DCT coefficient in the high frequency direction is further mirror-reflected to a higher frequency side to generate a third image.
5. A signal component up to the Nyquist frequency after the resampling is added by adding
3. The television receiver according to item 1. 6. The above-mentioned band limiting filter has a frequency response on the vertical axis on the y axis, a frequency on the horizontal axis on the x axis, a response at the origin of the frequency axis as A, and a Nyquist frequency after resampling as fn. 5. The television receiver according to claim 4, wherein y = AA * (x / fn) ^1.8 . 7. The method according to claim 1, wherein the band limiting filter is a panning window,
5. The television receiver according to claim 4, wherein the processing is performed using a window function including a Blackman window. 8. The method according to claim 1, wherein the second conversion processing means inserts a zero value into a high frequency side of the first signal component after adjusting a gain according to the magnification ratio. Item 4
3. The television receiver according to item 1. 9. The division means for dividing each DCT coefficient block into four sub-bands divided into a high-frequency component and a low-frequency component in the horizontal and vertical directions, respectively.
It is determined whether or not a component larger than a preset threshold value is present in the DCT coefficients of the subbands other than the lowband subband in both the horizontal and vertical directions among the four subbands divided by the division unit. The television receiver according to claim 4, further comprising: a component determination unit that performs the operation. 10. An image processing apparatus comprising: a DCT unit for performing DCT on an input image signal in a time-space domain; an edge detecting unit for detecting an edge in block units using a DCT coefficient DCT performed by the DCT unit; Image processing means for performing different enlargement processing for a block detected as a block and a block not including an edge; Inverse DCT means for performing inverse DCT with a basis matrix, and receiving edge information for each block from the edge detection means, and controlling the DCT means, the image processing means and the inverse DCT means for image resolution input from outside and output image resolution. Characterized by having system control means for controlling based on the included information Large equipment. 11. A first conversion processing unit that performs a pixel number conversion process on a block detected to include the edge among the two types of pixel number conversion processes performed by the image processing unit; And a second conversion processing unit that performs a pixel number conversion process on a block that is detected not to include the number of pixels. The first conversion processing unit includes a gain adjustment unit that adjusts a gain according to an enlargement ratio. After the original signal having the first signal component is mirror-folded in the frequency direction at the Nyquist frequency before resampling, the DCT coefficient of 1
Signal generating means for generating a second signal component shifted by a component in the high-frequency direction; mirror image inverting means for mirror-inverting the second signal component so that the positive or negative frequency response is inverted; and Nyquist after the resampling. Filter means for performing filter processing using a band-limiting filter up to a frequency, wherein a DCT coefficient of a high-frequency side of the pseudo high-frequency component generated by the gain adjustment means, the signal generation means, the mirror image inversion means, and the filter means is provided. 11. The method according to claim 10, wherein:
An image enlargement device according to claim 1. 12. The above-mentioned band-limiting filter has a frequency response on the vertical axis on the y-axis, a frequency on the horizontal axis on the x-axis, a response at the origin of the frequency axis as A, and a Nyquist frequency after resampling as fn. The image enlargement device according to claim 10, characterized in that: y = AA * (x / fn) ^1.8 . 13. The apparatus according to claim 10, wherein the band-limiting filter is performed using a window function including a panning window and a Blackman window. 14. The image processing device according to claim 2, wherein
Among the types of pixel number conversion processing, the pixel number conversion processing inserts a 0 value into the high frequency side of the original signal component after adjusting the gain according to the enlargement ratio for a block detected as not including an edge. The image enlargement apparatus according to claim 10, wherein the image enlargement is performed. 15. The division means for dividing each DCT coefficient block into four sub-bands divided into a high-frequency component and a low-frequency component in horizontal and vertical directions, respectively. And determining means for determining whether or not there is a component larger than a preset threshold in DCT coefficients of subbands other than the low-band subband in both the horizontal and vertical subbands among the divided subbands. The image enlargement device according to claim 10, wherein: 16. After the first signal component, which is the frequency component of the input original signal, is mirror-reflected in the frequency direction at the boundary of the Nyquist frequency before resampling, only one DCT coefficient component in the high frequency direction Folding step for generating a second signal component shifted to, an inversion step for inverting the second signal component to a mirror image and generating a third signal generation amount, and a third signal component generated by the inversion step Is added as the high frequency component of the input original signal to generate a pseudo high frequency component, and the Nyquist after the resampling from the low frequency component of the first signal component A filtering step of filtering frequency components up to a frequency band. 17. An insertion step of additionally inserting a pseudo high frequency component up to the Nyquist frequency after subsampling the first signal component, which is a frequency component of the input original signal, on the high frequency side of the original signal. Filtering the frequency components from the low-frequency component of the first signal component to the pseudo high-frequency component inserted in the inserting step. 18. A variable length decoding step of performing variable length decoding on an encoded bit stream, and an inverse quantization step of inversely quantizing a quantized DCT coefficient of a predetermined block decoded by the variable length decoding step. An edge detection step of detecting an edge in block units using the quantized DCT coefficient restored by the inverse quantization step; a first block detected to include an edge by the edge detection step; An image processing step of performing different enlargement processing with the second block detected not to be included, and a base matrix corresponding to the enlarged block size for the DCT coefficient block subjected to the image processing by the image processing step Performs an inverse DCT step of performing an inverse DCT and local decoding by motion compensation in consideration of the conversion of the number of pixels. A decoding step of decoding the video signal before being converted into a video signal, and a pixel number conversion step of converting the number of pixels of the video signal processed by the decoder step into an appropriate number of pixels to be displayed on the display means. A system control step of controlling the content of processing performed in the decoder step based on information including image resolution and edge information of the video signal processed in the decoder step; and the decoder step And a display control step of controlling the display process in the display step, while converting the signal processed in the display step into a video signal displayable in the display step. 19. The image processing step includes a first conversion step of performing a pixel number conversion process on a block detected as including the edge, and a pixel conversion process on a block detected as not including the edge. A second conversion step of performing a number conversion process, wherein the first conversion step includes a gain adjustment step of adjusting a gain according to an enlargement ratio, and a re-sampling of the original signal having the first signal component. A signal generating step of generating a second signal component shifted in the high frequency direction by one component of the DCT coefficient after performing mirror image folding in the frequency direction with the Nyquist frequency as a boundary, and converting the second signal component into a frequency response A mirror image inversion step of inverting the mirror image so that the sign of the image is inverted, and a filter step of performing a filtering process using a band-limiting filter up to the Nyquist frequency after the resampling. And, according to claim 18, characterized by using the gain adjustment step, the signal generation step, mirror-reversing step, and a pseudo high frequency component which has been treated by filter process as DCT coefficients of the high frequency side
3. The method for processing a television signal according to item 1. 20. The signal generation step, wherein the second signal component is a component up to the Nyquist frequency after the resampling when the enlargement ratio is smaller than 2, and the enlargement ratio is larger than 2 times. In this case, after the mirror image is turned back in the frequency direction, a component shifted by one component of the DCT coefficient in the high frequency direction is further mirrored back to the higher frequency side, and a third signal component generated is added. 20. The method for processing a television signal according to claim 19, wherein the component is a component up to the Nyquist frequency. 21. The characteristics of the band limiting filter are as follows: the frequency response on the vertical axis is on the y-axis, the frequency on the horizontal axis is on the x-axis, the response at the origin of the frequency axis is A, and the Nyquist frequency after resampling is fn. 20. The television signal processing method according to claim 19, wherein y = AA * (x / fn) ^1.8 . 22. The method according to claim 19, wherein the band limiting filter is performed using a window function including a panning window and a Blackman window. 23. The method according to claim 23, wherein in the second conversion step, after adjusting a gain according to the enlargement ratio, a zero value is inserted in a high frequency side of the first signal component. 19
3. The method for processing a television signal according to item 1. 24. The edge detecting step includes a dividing step of dividing each DCT coefficient block into four subbands divided into a high frequency component and a low frequency component in the horizontal and vertical directions, respectively. A component determining step of determining whether or not a DCT coefficient of a sub-band other than the sub-band in which both the horizontal and vertical bands are low in the four sub-bands includes a component larger than a predetermined threshold value; 20. The method of processing a television signal according to claim 19, comprising: 25. A DCT step for DCT of an input image signal in a time-space domain, an edge detection step for detecting an edge in a block unit using a DCT coefficient DCT performed by the DCT step, and an edge detected by the edge detection step. An image processing step of performing different enlargement processing on a block detected as a block and a block not including an edge, and adjusting the DCT coefficient block after the enlargement processing in the image processing step to the enlarged block size. An inverse DCT step of performing an inverse DCT on a basis matrix; and receiving edge information for each block from the edge detection step, and performing the DCT step, the image processing step, and the inverse DCT step on an image resolution and an output image resolution input from outside. Comprising a system control step of controlling based on information including Large way. 26. A first conversion step of performing a pixel number conversion process on a block detected to include the edge among the two types of pixel number conversion processes performed in the image processing step; A second conversion step of performing a pixel number conversion process on a block detected not to be included, wherein the first conversion step includes a gain adjustment step of adjusting a gain according to an enlargement ratio; A signal generation step of performing a mirror image folding back of the original signal having the signal component of the above at the Nyquist frequency before resampling in the frequency direction to generate a second signal component; A mirror image inverting step of inverting a mirror image so that is inverted, and a filter step of performing a filtering process using a band-limiting filter up to the Nyquist frequency after the resampling, Gain adjustment step, according to claim, characterized by using signal generating step, mirror-reversing step, and a pseudo high frequency components generated by the filter process as DCT coefficients of the high frequency side 25
The image enlargement method described in 1. 27. The band limiting filter process according to claim 27, wherein the frequency response on the vertical axis is on the y-axis and the frequency on the horizontal axis is on the x-axis.
26. The processing according to the characteristic given by the following equation, y = AA * (x / fn) ^1.8, where A is a response at the origin of the frequency axis and fn is a Nyquist frequency after resampling. The described image enlargement method. 28. The image enlarging method according to claim 25, wherein said band limiting filter processing is performed using a window function including a panning window and a Blackman window. 29. A method according to claim 29, wherein
Among the types of pixel number conversion processing, the pixel number conversion processing inserts a value of 0 into the high-frequency side of the original signal component after adjusting the gain according to the enlargement ratio for a block detected as not including an edge. The image enlargement method according to claim 25, wherein the image enlargement method is performed. 30. The edge detecting step includes dividing each DCT coefficient block into four sub-bands divided into a high-frequency component and a low-frequency component in the horizontal and vertical directions, respectively. A determination step of determining whether there is a component larger than a preset threshold value in the DCT coefficients of the subbands other than the subband that is low in both the horizontal and vertical directions among the two divided subbands. The image enlargement method according to claim 25, wherein the method is performed. 31. A storage medium storing a program constituting each means according to claim 1 so as to be readable from a computer. 32. A storage medium storing a program for executing the method according to claim 16 in a manner readable from a computer.