JPH08307834A - Image information converter and its method - Google Patents

Abstract

PURPOSE: To convert an SD image into an HD image in one mode. CONSTITUTION: A changeover device 2 receiving an SD image gives an odd numbered field to a delay circuit 5 in response to a field ID received from a terminal 3 and gives an even numbered field to a phase shift filter circuit 4. A re-quantization value is fed from an area division circuit 6 to a class code generating circuit 11 via an ADRC circuit 9 and a class code generating circuit 11 receives a motion class mv-class via a motion class decision circuit 8. A class code class from the class code generating circuit 11 is fed to a ROM table 12, from which coefficient data wi(class) are fed from the ROM table 12 to an estimate arithmetic circuit 13. The estimate arithmetic circuit 13 generates an HD image via a horizontal interpolation filter 14.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image information conversion device suitable for use in, for example, a television receiver, a video tape recorder device, or the like, and particularly to a high resolution image information of ordinary resolution supplied from the outside. The present invention relates to an image information conversion device that converts and outputs image information.

[0002]

2. Description of the Related Art Nowadays, due to the increasing audio-visual orientation, it is desired to develop a television receiver capable of obtaining a higher resolution image. In response to this demand, a so-called high vision has been developed. . The number of scanning lines defined in the so-called NTSC system is 525 in this high-definition system, which is more than doubled to 1125 lines, and the aspect ratio of the display screen is 9: 3 in the NTSC system 3: 4. 16 and wide-angle screen. Therefore, it is possible to obtain a screen with high resolution and a realistic feeling.

Here, although it is a high-definition television having such excellent characteristics, it is not possible to display an image even if an NTSC video signal is supplied as it is. This is because the standards are different between the NTSC system and the high-definition system as described above. For this reason, when an image corresponding to an NTSC video signal is to be displayed in high-definition, the video signal rate conversion has conventionally been performed using an image information conversion device as shown in FIG. 8, for example.

In FIG. 8, a conventional image information conversion apparatus includes a horizontal interpolation filter 101 for performing horizontal interpolation processing on an NTSC video signal supplied via an input terminal 100, and a horizontal interpolation processing. And a vertical interpolation filter 102 that performs interpolation processing in the vertical direction on the performed video signal. Then, the output terminal 103
It is possible to obtain a high definition video signal from.

Specifically, the horizontal interpolation filter 101 is
The video signal of the NTSC system having the configuration as shown in FIG. 9 is supplied via the input terminal 100 to the first to mth multipliers 111 to 111 _m via the input terminal 110, respectively. To be done. Each multiplier 111 multiplies the video signal by a coefficient and outputs the result. The video signals multiplied by the coefficients are added to the first to mth adders 112 to 112, respectively.
supplied to _m-1 . Delay registers 113 to 113 _{m for} time T are provided between the adders 112 to 112 _m-1 , respectively. Then, the video signal output from the m-th multiplier 111 _m is delayed by the time T by the m-th delay register 113 _m , and the m-th adder 112 _m-1
Is supplied to.

The (m-1) th adder 112 _m-1 outputs from the _mth delay register 113 _m the video signal delayed by the time T and the (m-1) th multiplier 111 _m-1. And the video signal of are added and output. The video signal on which this addition processing has been performed is again subjected to the delay time of time T by the (m−1) th delay register 113 _m−1 , and the (m−2) th not shown
Of the adder 112 _{m-2 of} FIG.
The video signal from the _second multiplier 112 _m-2 is added. In this way, the horizontal interpolation filter 101
The video signal of the C system is supplied to the vertical interpolation filter 102 via the output terminal 120.

The vertical interpolation filter 102 has the same structure as the horizontal interpolation filter 101 described above, and performs vertical pixel interpolation on the video signal subjected to the horizontal interpolation processing. As a result, vertical pixel interpolation is performed on the NTSC video signal. The high-definition video signal thus converted is supplied to the high-definition image receiver. As a result, an image corresponding to the NTSC video signal can be displayed on the high-definition receiver.

However, the conventional image information conversion apparatus merely interpolates in the horizontal and vertical directions based on the video signal of the NTSC system, so that the resolution is based on the video of the base NTSC system. It was no different from the signal. In particular, when a normal moving image is to be converted, it is common to perform vertical interpolation by intra-field processing, but in that case, since the inter-field correlation of the image is not used, in the image static part, There is a drawback that the resolution is deteriorated rather than the NTSC video signal.

On the other hand, the applicant has filed Japanese Patent Application No. 6-205
No. 934, the image signal conversion apparatus performs storage in accordance with a three-dimensional (spatio-temporal) distribution of image signal levels as input signals, and stores a prediction coefficient value acquired in advance for each class by learning. It has been proposed that an optimum estimated value be output by a calculation based on a prediction formula.

According to this method, when an HD pixel is created, a plurality of SD pixel data in the vicinity of the HD pixel to be created is used for class division, and a prediction coefficient value is acquired by learning for each class. This is a subtle technique in which the HD pixel value closer to the true value is obtained by utilizing the intra-frame correlation in the image static portion and the intra-field correlation in the moving portion.

According to this method, the switching between still / moving can be smoothly expressed by learning by using an actual image, so that it is unnatural due to the switching between still / moving as in the conventional motion adaptive method. It is possible to significantly reduce the occurrence of the sag.

[0012]

However, the method disclosed in the above-mentioned prior application has a pixel structure as shown in FIG. 10 in order to equalize the intervals of HD pixels to be created. That is, it is necessary to have two modes: a mode for creating an HD pixel y ₁ close to the central SD pixel x and a mode for creating an HD pixel y ₂ far from the SD pixel x. Since it is impossible to share a coefficient between these two modes, two sets of coefficients, a coefficient for mode 1 and a coefficient for mode 2, are required for each class classified. By the way, in order to improve the accuracy of conversion, it is necessary to have many classes. Therefore, having a plurality of modes means that the coefficient ROM becomes large at that rate, which is a problem in practical use.

Further, in mode 1 for producing HD pixels close to SD pixels, high-precision conversion is always possible, but in mode 2 for producing HD pixels far from SD pixels, other fields are used in the static part. Highly accurate conversion is possible because the spatially close data can be used for conversion. However, in the case of a moving part, the data cannot be used for conversion, so the conversion accuracy deteriorates. This causes a considerable difference in conversion accuracy between the mode 1 and the mode 2 in the moving image part, and sometimes causes a problem in image quality.

Now, in the vertical conversion of the separable system in which the conversion in the vertical direction and the conversion in the horizontal direction are performed in order (not simultaneously), two conversion modes are necessary because the HD pixel interval created as described above is used. This is to make it correct. In order to make the produced HD pixel intervals uniform, two modes with unequal distances from SD pixels are required. Here, assuming that the HD pixels are created so that the distances from the SD pixels are equal so that two modes are not prepared, the pixel relationship is as shown in FIG. That is, the HD pixels are spatially located at the same position in the first field and the second field, and the HD pixel interval is incorrect. Therefore, this method cannot be used simply.

In order to correct the HD pixel interval by using the above method, it is necessary to change the processing in the first field and the second field. For example, when the pixel structure shown in FIG. 12 is used, the HD pixel interval becomes correct. However, only one mode is required for the first field, but two modes different from the first field are required for the second field, which means that a total of three modes are required.
Not only the reduction of the coefficient ROM but also the increase thereof.
As described above, in the conventional approach, it is difficult to reduce the ROM by reducing the modes.

By the way, most of the conventional conversion systems are separable systems. For example, when the conversion in the vertical direction is 2 modes and the conversion in the horizontal direction is 1 mode, the separable method requires coefficients for 3 modes in total in (2 + 1). In the non-separable method in which the conversions of (1) and (2) are performed simultaneously, (2 × 1) is sufficient for a total of two modes. Therefore, although it may be possible to reduce the number of modes by adopting the non-separable method, when trying to obtain equivalent performance,
Since the number of classes in the class classification increases, the coefficient R is substantially
It does not lead to the reduction of OM. Therefore, the coefficient ROM
From the perspective of reducing emissions, the adoption of the non-separable method is not the deciding factor.

The present invention has been made in view of the above problems, and an object of the present invention is to provide an image information conversion apparatus capable of converting an NTSC video signal into a high definition video signal with improved resolution. And

[0018]

According to a first aspect of the present invention, in an image information converting apparatus for converting a digital image signal into a digital image signal having a larger number of pixels, image information supplied from the outside is supplied. Switching means for switching between the odd field and the even field for output, a delay means for delaying and outputting the odd field, and a phase shift means for phase-shifting and outputting the even field,
An image information division unit that divides the data supplied from the delay unit and the phase shift unit into a plurality of blocks composed of a plurality of image data located temporally and spatially, and an image for each block divided by the image information division unit. A pattern of the level distribution of information is detected, and based on the detected pattern, class detection means for outputting class information to which the image information of the block belongs, and image information supplied from the outside, image information supplied from the outside Coefficient data of an estimation formula, which is information for converting to image information of higher resolution, is stored for each class, and coefficient data storage means for outputting coefficient data according to class information from the class detection means, According to the coefficient data supplied from the coefficient data storage means, the image information supplied from the outside is changed from the image information supplied from the outside. Also image information converting apparatus, comprising an image converting means for converting the high-resolution image information.

According to a third aspect of the present invention, in an image information conversion method for converting a digital image signal into a digital image signal having a larger number of pixels, an odd field of image information supplied from the outside is used. Switching even-numbered fields and outputting them, delaying odd-numbered fields and outputting them, outputting even-fields by phase-shifting them, and outputting the data supplied from the delay and shift steps spatially and spatially A step of dividing a plurality of image information located in a plurality of blocks into a plurality of blocks, a pattern of level distribution of the image information is detected for each divided block, and based on the detected pattern, class information to which the image information of the block belongs And the image information supplied from the outside, the image supplied from the outside. The coefficient data of the estimation formula, which is the information for converting into the image information of the resolution higher than that of the report, is stored for each class, and the step of outputting the coefficient data according to the class information and the coefficient data supplied And converting the image information supplied from the outside into image information having a resolution higher than that of the image information supplied from the outside and outputting the image information.

[0020]

In the image information converting apparatus according to the present invention, when the odd field data is generated, the input SD signal is sent to the image information dividing means as it is, and when the even field data is generated, the input SD signal is phased. The shift information is shifted up by 1/4 of the SD pixel interval in the field and sent to the image information dividing means. The image information dividing means divides into a plurality of regions, each of which is composed of a plurality of pixels in the same frame that are continuous in the vertical direction, detects a pattern of the level distribution of the image information for each region, and based on the detected pattern. Then, the class to which the image information of the area belongs is determined and the class detection information is output. further,
Dividing into a combination of a plurality of inter-frame data by different types of image information dividing means, calculating an average absolute value of the inter-frame difference at the spatially same position for each area, and setting a preset threshold. A class indicating the degree of movement is determined by the value, and class detection information is output. The two classes are integrated by the class code generation circuit and output as the final class. The coefficient data storage means stores, for each class, coefficient data of a linear estimation formula which is information for converting image information supplied from the outside into image information having a resolution higher than this image information. The data is output according to the class detection information. Then, the image information conversion means converts the image information supplied from the outside into image information having a higher resolution than the image information supplied from the outside according to the coefficient data supplied from the coefficient data storage means.

[0021]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The embodiments of the image signal converting apparatus according to the present invention will be described below in detail with reference to the drawings. FIG.
Shows a schematic configuration of the signal processing of this embodiment, that is, the image signal conversion apparatus. Image information externally supplied from the input terminal 1 is, for example, so-called NTSC.
The video signal of the system is digitized, and SD (Standerd D
efinition) data.

The positional relationship between the SD pixel and the HD pixel to be created in this embodiment is as shown in FIG. 3 in the first field. That is, H to be created
When viewed in the same field, the D pixels are present at equal distances above and below the SD pixels. That is, the central SD
HD pixel y ₁ located above pixel x, H located below
The D pixel y ₂ is at the same distance from the SD pixel x. Hereinafter, the mode for estimating the HD pixel existing at the position above the SD pixel will be referred to as mode 1, and the mode for estimating the HD pixel existing at the position below the SD pixel will be referred to as mode 2.

However, when the same structure is used to create the HD pixels of the second field, the HD pixels of the first field and the second field are spatially created at the same position as described above. This method cannot be used.

Therefore, in creating the HD image of the first field, the SD image data is treated as it is, and in creating the HD image of the second field, the SD image subjected to the phase shift by filtering is treated.

The SD data supplied from the input terminal 1 is
The switch 2 is supplied to the delay circuit 5 if the SD data supplied from the input terminal 1 is the SD data of the first field, based on the field ID supplied from the terminal 3. If it is the SD data of the second field, it is supplied to the phase shift filter circuit 4.
The phase shift filter circuit 4 is a circuit that shifts the phase of SD pixels for generating HD pixels in the second field. Here, all SD pixels are 1 / of the intra-field SD pixel interval by the intra-field phase shift filter.
The phase is shifted to a position shifted up by 4. The phase shift filter is not particularly limited, but a filter close to an ideal filter is desirable from the viewpoint of conversion performance. As a result, new SD pixel data is generated at the position shown in FIG. In creating the HD image of the second field, HD image data is created by using these SD images generated by filtering in the same manner as in the first field. That is, HD pixels are generated at the upper and lower positions of the SD image data generated by filtering. This state is shown in FIG.

As described above, with respect to the HD pixels of the first field, the input SD pixel data is treated as it is, and with respect to the HD pixels of the second field, the phase-shifted SD pixel data is treated to generate the HD image. Thus, it is possible to generate HD images at regular intervals in one mode.

On the other hand, the delay circuit 5 uses the H of the first field.
This is a circuit that delays the same amount of time as the phase shift filter circuit 4 requires to generate D pixels. When the HD pixel of the first field is generated, the output signal of the delay circuit 5 is supplied to the area dividing circuit 6 and the area dividing circuit 7,
Similarly, when the HD pixel of the second field is generated, the output signal of the phase shift filter circuit 4 is supplied to the area dividing circuit 6 and the area dividing circuit 7.

For simplicity, the following description will focus on the case of HD pixel generation in the first field. In the area dividing circuit 6,
The SD image signal supplied from the phase shift filter circuit 4 or the delay circuit 5 is divided into a plurality of areas. In this embodiment, from the SD pixel belonging to the same field of the HD pixel to be created and the SD pixel belonging to the immediately preceding field, for example, from the spatially adjacent pixels in the vertical direction, 5
One pixel is selected and divided into an area consisting of 1 pixel × 5 lines and a total of 5 pixels.

SD pixels x ₁ , x ₂ , x ₃ , x ₄ , x ₅ for the HD pixel y ₁ , HD pixel y ₂ in FIG.
Is in that area.

The data divided into blocks by the area dividing circuit 6 is supplied to the ADRC circuit 9 and the delay circuit 10. The delay circuit 10 delays the data for the time required for the processing of the ADRC circuit 9, the class code generation circuit 11, and the ROM table 12, and outputs the delayed data to the estimation calculation circuit 13.

The ADRC circuit 9 detects the pattern of the one-dimensional or two-dimensional level distribution of the SD data supplied for each area as described above, and the data of each area as described above, for example, 8 2 from bit SD data
Pattern compression data is formed by performing an operation such as compression into bit SD data, and this pattern compression data is supplied to the class code generation circuit 11. Originally, A
DRC (Adaptive Dynamic Range Coding) is a VTR
Although it is an adaptive requantization method developed for high-efficiency coding for digital signals, it can efficiently express a local pattern of signal levels with a short word length. Used for code generation. The ADRC circuit 9 sets the maximum value MA in the area by using the following equation (1) where DR is the dynamic range in the area, n is the bit allocation, L is the data level of the pixels in the area, and Q is the requantization code.
The region between X and the minimum value MIN is equally divided by the designated bit length and requantization is performed.

DR = MAX-MIN + 1 Q = [(L-MIN + 0.5) · 2 ⁿ / DR] (1) where [] means rounding down.

In this embodiment, it is assumed that the SD data of 5 pixels separated by the area separation circuit 6 is compressed into 2 bits. Let the compressed SD data be q ₁ , q ₂ , q ₃ , q ₄ , and q ₅ , respectively.

On the other hand, the SD image signal supplied from the phase shift filter circuit 4 or the delay circuit 5 is also supplied to the area dividing circuit 7. Also in the area dividing circuit 7,
The supplied SD image signal is divided into a plurality of areas. In this embodiment, from the supplied SD image signal, the SD pixel two fields before the HD pixel to be created is divided into an area consisting of a total of 3 pixels, for example, 1 pixel × 3 lines, and the HD pixel to be created further. The SD pixels in the same field as the above are similarly divided into areas each having a total of 3 pixels of 1 pixel × 3 lines.

That is, the HD pixels y ₁ and H in FIG.
SD pixel of the previous frame for the D pixel y ₂ m _1, m _2,
m ₃ and SD pixels n ₁ , n ₂ , and n _{3 of the} same frame correspond to the area.

The data cut out by the area dividing circuit 7 is supplied to the motion class determining circuit 8. The motion class determination circuit 8 supplies the SD supplied for each area as described above.
The difference between the data is calculated, and the average value of the absolute values is thresholded to calculate a motion parameter that is a motion index, and this motion class mv-class is supplied to the class code generation circuit 11. Specifically, the motion class determination circuit 8
Calculates the average value param of the absolute values of the differences of the supplied SD data by the following equation (2).

[0037]

[Equation 1] However, in this embodiment, n = 3.

The average value param of the absolute values of the SD data differences calculated by the above method is set by a preset threshold value,
The motion class mv-class is calculated using the average value param of the absolute values of the differences of the SD data. For example, here, four motion classes are provided, and the motion class mv-class is determined as follows. If param ≤2: Motion class 0 If param ≤4: Motion class 1 If param ≤8: Motion class 2 If param> 8: Motion class 3

The class code generation circuit 11 receives the pattern compression data (q ₁ , q ₂ ,
q ₃ , q ₄ , q ₅ ) and the motion class mv-class supplied from the motion class determination circuit 8 based on the following equation (3)
The class to which the block belongs is detected and the class code class indicating that class is RO
Supply to the M table 12. This class code class
Indicates the read address from the ROM table 12.

[0040]

[Equation 2]

In this embodiment, n = 3 and P = 2.
However, since the pixel positions are opposite in the estimation in mode 1 and the estimation in mode 2, it is necessary to reversely read the ADRC quantized data in modes 1 and 2. That is, when the class of mode 1 is determined by the equation (2), the class of mode 2 is determined by the following equation (4).

[0042]

(Equation 3)

In the ROM table 12, coefficient data for calculating HD data corresponding to SD data is learned for each class by using a linear estimation formula by learning the relationship between SD data patterns and HD data. Remembered This is image information having a higher resolution than the image information of the SD data by the linear estimation formula. This is information for converting into HD (High Definition) data that conforms to the so-called high-definition standard. In this example,
Coefficient data is prepared commonly for Mode 1 and Mode 2.
The method of creating the coefficient data stored in the ROM table 12 will be described later. From the ROM table 12, the coefficient data wi (class) of the class is read from the address indicated by the class code class. This coefficient data is supplied to the estimation calculation circuit 13.

The estimation operation circuit 13 receives the SD data and R supplied from the area dividing circuit 6 via the delay circuit 10.
Wi which is the coefficient data supplied from the OM table 12
HD data corresponding to the input SD data is calculated based on (class).

More specifically, the estimation operation circuit 13 uses the SD data supplied from the delay circuit 10 by the coefficient data supplied from the ROM table 12 based on wi (class) which is the coefficient data. The HD data corresponding to the input SD data is calculated by performing the calculation shown in the equation (5). The created HD data is
It is supplied to the horizontal interpolation filter 14.

Hd ′ = w ₁ x ₁ + w ₂ x ₂ + w ₃ x ₃ + w ₄ x ₄ + w ₅ x ₅ (5)

The horizontal interpolation filter 14 is the same as the horizontal interpolation filter 102 of FIG. 8, and doubles the number of pixels in the horizontal direction by the interpolation processing. The output of the horizontal interpolation filter 14 is output via the output terminal 15. The HD data output via the output terminal 15 is supplied to, for example, an HD television receiver, an HD video tape recorder device, or the like.

As described above, the coefficient data for estimating the HD data corresponding to the SD data is obtained by learning for each class in advance, and is stored in the ROM table 12, and the SD data and the ROM to be input are input. The input SD data is simply interpolated by performing calculation based on the coefficient data (wi (class)) read from the table 12 and forming and outputting HD data corresponding to the input SD data. Unlike, it is possible to output data closer to the actual HD data.

Next, a method of creating coefficient data stored in the ROM table 12 will be described with reference to FIG.

To obtain coefficient data by learning,
First, an SD image corresponding to an already known HD image and having a pixel number of ¼ of the HD image is formed. In particular,
With the ideal filter circuit shown in FIG. 2, the pixels in the vertical direction of the HD data supplied through the input terminal 21 are thinned by the vertical thinning filter 22 so that the vertical frequency in the field becomes 1/2. Further, in the horizontal thinning filter 23, SD data is obtained by thinning out pixels in the horizontal direction of the HD data. The SD data obtained by the vertical thinning filter 23 is supplied to the area dividing circuit 24.

In the area dividing circuit 24, the SD image signal supplied from the horizontal thinning filter 23 is divided into a plurality of areas. Specifically, the area dividing circuit 24 has the same function as that of the area dividing circuit 6 described above. In this embodiment, like the area dividing circuit 6, each 5
The SD image signal is divided into areas including pixels. The SD data for each area is supplied to the ADRC circuit 25 and the normal equation adding circuit 29.

The ADRC circuit 25 detects a one-dimensional or two-dimensional level distribution pattern of the SD data supplied for each area and, as described above, all or part of the data in each area. , For example, 8-bit S
Pattern compression data is formed by performing an operation such as compression from D data to 2-bit SD data, and this pattern compression data is supplied to the class code generation circuit 28. This ADRC circuit 25 corresponds to the AD described above.
This is the same as the RC circuit 9, and in this embodiment, the SD data (x _{1 to} x ₅ in FIG. 6) of each area consisting of 5 pixels separated by the area dividing circuit 24 is ADRC.
To compress each to 2 bits.

On the other hand, the SD image signal supplied from the horizontal thinning filter 23 is also supplied to the area dividing circuit 26. Specifically, the area dividing circuit 26 has the same function as that of the area dividing circuit 7 described above. The SD data cut out by the area dividing circuit 26 is supplied to the motion class determining circuit 27. Specifically, the motion class determination circuit 27 has the same function as the motion class determination circuit 8 described above. The motion class determined by the motion class determination circuit 27 is supplied to the class code generation circuit 28.

The class code generation circuit 28 is the same as the class code generation circuit 11 described above, and AD
The class to which the block belongs is detected by performing the operation of Expression (2) based on the pattern compressed data supplied from the RC circuit 25 and the motion class supplied from the motion class determination circuit 27, and a class indicating that class. It outputs the code. Class code generation circuit 28
Outputs the class code to the normal equation adding circuit 29.

Here, in order to explain the normal equation adding circuit 29, learning of a conversion formula from a plurality of SD pixels to HD pixels and signal conversion using the prediction formula will be described. For the sake of explanation, a case will be described below in which pixels are generalized and prediction is performed using n pixels. Further, the SD pixel levels are set to x ₁ , ..., X _n , respectively, and the requantized data resulting from p-bit ADRC is q
_{Let 1} , ..., q _n .

At this time, the class code class of this area
Is defined by equation (2).

As described above, when the SD pixel level is x ₁ , ..., X _n and the HD pixel level is y, n taps by the coefficients w ₁ , ..., W _n for each class. Set the linear estimation formula of. This is shown in equation (6).
Before learning, w _i is an undetermined coefficient.

Y = w ₁ x ₁ + w ₂ x ₂ + ... + w _n x _n (6)

Learning is performed on a plurality of signal data for each class. When the number of data is m, the following equation (7) is set according to the equation (6).

Y _k = w ₁ x _k1 + w ₂ x _k2 + ... + w _n x _kn (7) (k = 1, 2, ... M)

When m> n, w ₁ , ..., W _n are not uniquely determined. Therefore, the elements of the error vector e are defined by the equation (8), and the coefficient that minimizes the equation (9) is set. Ask. This is the so-called least squares method.

E _k = y _k − {w ₁ x _k1 + w ₂ x _k2 + ... + w _n x _kn } (8) (k = 1, 2, ..., m)

[0063]

[Equation 4]

Here, the partial differential coefficient by w _{i in} equation (9) is obtained. It makes each w _i such that equation (10) becomes zero.
You should ask.

[0065]

(Equation 5)

Below, as in equations (11) and (12),
Defining X _ij Y _i , equation (10) can be rewritten as equation (13) using a matrix.

[0067]

(Equation 6)

[0068]

(Equation 7)

[0069]

(Equation 8)

This equation is generally called a normal equation. The normal equation addition circuit 29 receives the SD code x ₁ supplied from the class code generation circuit 28, the SD data x ₁ , ..., X _n supplied from the area dividing circuit 24 from the input terminal 21. HD corresponding to
This normal equation is added using the pixel level y.

After inputting all the training data, the normal equation adding circuit 29 outputs the normal equation data to the prediction coefficient determining circuit 30. The prediction coefficient determination circuit 30 solves w _i using a general matrix solution method such as a sweeping method of a normal equation and calculates a prediction coefficient.
The prediction coefficient determination circuit 30 writes the calculated prediction coefficient in the memory 31.

As a result of training as described above,
The memory 31 stores the quantized data q ₁ , q ₂ , q ₃ ,
For each pattern defined by q ₄ and q ₅ , a prediction coefficient for estimating the target HD data y is stored that is statistically closest to the true value. The table stored in the memory 31 is the ROM table 12 used in the image signal converting apparatus of the present invention as described above.
By the above processing, H is calculated from SD data by the linear estimation formula.
The learning of the coefficient data for creating the D data ends.

In the above description of the embodiment, ADRC is provided as the information compression means, but this is just an example, and any information compression means that can be expressed in a class having a small number of signal waveform patterns can be used. What is provided is arbitrary, and compression means such as DPCM or VQ may be used.

Further, in the description of the above-mentioned embodiment, the horizontal up-conversion and the horizontal interpolation filter 14 are used for the sake of simplicity. Instead, however, a ROM for the horizontal up-conversion is prepared, and the horizontal up-conversion is prepared. It is of course possible to adopt the method using the estimation formula also in the up-conversion of.

Further, in the description of the above embodiment, the separable system in which the conversion in the vertical direction and the conversion in the horizontal direction are sequentially performed is adopted, but this is not the essence of the present invention, and the conversion in the vertical direction and the horizontal direction are performed. There is no problem even if a non-separable method is adopted in which the conversion of the above is performed at the same time.

Further, in the above description of the embodiment, the pattern of the signal waveform is one-dimensionally divided by the area dividing circuit 6, but it may be two-dimensionally divided.

Furthermore, in the above description of the embodiment, the motion class mv-c is calculated by using the SD image data which is one-dimensionally divided by the area dividing circuit 7 and the motion class determining circuit 8.
Although the lass is determined, the area division may be two-dimensional division. Rather, it is preferable to make it two-dimensional. Further, for simplicity, this time, the area division by the area division circuit 6 and the area division by the area division circuit 7 are similar area divisions, but these are originally completely different, and similar area divisions are performed. There is no need to do it.

Further, in the above description of the embodiment, the SD pixel used for class classification and the SD pixel used in the linear estimation formula are the same, but they do not necessarily have to be the same. When different pixels are used, it is desirable that the SD pixels used in the class classification be included in the SD pixels used in the linear estimation formula, and the SD pixels used in the linear estimation formula additionally used are: Estimate H
It is desirable that only the pixels belonging to the same field as the D pixel are included.

Further, in the description of the above embodiment, the ROM
When creating the table 12, only SD data that is not phase-shifted is input, but as shown in FIG. 1, circuits similar to the phase shift filter circuit 4 and the delay circuit 5 are prepared, and the SD data that has been phase-shifted is prepared. May be a learning target.

[0080]

According to the present invention, the processing is changed between the odd field and the even field, and the HD pixel generation in the even field is performed by shifting the phase of the SD by the phase shift filter.
By using the pixel-based method, vertical conversion is realized in one conversion mode. As a result, the size of the ROM table is halved, and the image quality can be homogenized.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of an image information conversion apparatus according to the present invention.

FIG. 2 is a block diagram for explaining generation of a correction data table.

FIG. 3 is a schematic diagram for explaining the phase relationship of the conversion system of the present invention.

FIG. 4 is a schematic diagram for explaining the phase relationship of the conversion system of the present invention.

FIG. 5 is a schematic diagram for explaining a problem of a conventional space-time class division method.

FIG. 6 is a schematic diagram for explaining data used for space-time class determination according to the present invention.

FIG. 7 is a schematic diagram for explaining data used for determining a motion class according to the present invention.

FIG. 8 is a block diagram of a conventional image information conversion device.

FIG. 9 is a circuit diagram of an example of a horizontal interpolation filter according to a conventional image information conversion device.

FIG. 10 is a schematic diagram for explaining a phase relationship of a conventional conversion method.

FIG. 11 is a schematic diagram for explaining a phase relationship of a conventional conversion method.

FIG. 12 is a schematic diagram for explaining a phase relationship of a conventional conversion method.

[Explanation of symbols]

 2 switcher 4 phase shift filter circuit 6, 7 area division circuit 8 motion class determination circuit 9 ADRC circuit 11 class code generation circuit 12 ROM table 13 estimation calculation circuit 14 horizontal interpolation filter

Claims (4)

[Claims] 1. An image information conversion device adapted to convert a digital image signal into a digital image signal having a larger number of pixels. Switching for switching between an odd field and an even field of image information supplied from the outside. Means, a delay means for delaying and outputting the odd field, a phase shift means for phase-shifting and outputting the even field, and a data supplied from the delay means and the phase shift means spatially and spatially close to each other. Image information dividing means for dividing the image information into a plurality of blocks composed of a plurality of image data, and a level distribution pattern of image information is detected for each of the blocks divided by the image information dividing means, and the detected pattern is detected. Based on, class detection means for outputting class information to which the image information of the block belongs, The coefficient data of the estimation formula, which is the information for converting the image information supplied from the outside to the image information having a higher resolution than the image information supplied from the outside, is stored for each class, and the class is stored. The coefficient data storage means for outputting the coefficient data according to the class information from the detection means, and the image information supplied from the outside according to the coefficient data supplied from the coefficient data storage means And an image conversion unit for converting and outputting image information having a higher resolution than the image information supplied from the image information conversion apparatus. 2. The image information conversion apparatus according to claim 1, wherein the coefficient data storage means has a memory means for storing the coefficient data for each class, and spatially and / or temporally of a pixel of interest. When the value of the pixel of interest is created by the linear linear combination of the values of a plurality of pixels in the vicinity and the coefficient data, for each class that minimizes the error between the created value and the true value of the pixel of interest. An image information conversion device, characterized in that the coefficient data is obtained in advance by learning. 3. A method of converting a digital image signal into a digital image signal having a larger number of pixels, wherein the step of outputting an odd field and an even field of image information supplied from the outside by switching A delay step for delaying and outputting the odd field, a shift step for phase-shifting and outputting the even field, and data provided from the delay step and the shift step are temporally and spatially close to each other. A step of dividing a plurality of image information into a plurality of blocks, a pattern of the level distribution of image information is detected for each of the divided blocks, and class information to which the image information of the block belongs based on the detected pattern And outputting the image information supplied from the outside Coefficient data of an estimation formula, which is information for converting into image information having a higher resolution than the image information supplied from the unit, is stored for each class, and the step of outputting the coefficient data according to the class information And converting the image information supplied from the outside into image information having a higher resolution than the image information supplied from the outside and outputting the image information according to the supplied coefficient data. Image information conversion method. 4. The image information conversion method according to claim 3, further comprising a step of storing the coefficient data for each class, and a value of a plurality of pixels spatially and / or temporally neighboring the pixel of interest. When the value of the pixel of interest is created by linear linear combination of the coefficient data, the coefficient data for each class that minimizes the error between the created value and the true value of the pixel of interest is obtained by learning in advance. A method for converting image information, which is characterized in that