JP4001143B2 - Coefficient generation apparatus and method - Google Patents

Description

  The present invention relates to an image information conversion apparatus suitable for use in, for example, a television receiver, a video tape recorder apparatus, and the like. The present invention relates to a coefficient generation apparatus and method for generating coefficients used when converting into image information.

  Nowadays, development of a television receiver capable of obtaining a higher resolution image is desired due to an increase in audio / visual orientation, and in response to this demand, a so-called high vision system has been developed. In this high-definition system, the number of scanning lines stipulated in the so-called NTSC system is 525, whereas it is 1125 more than twice, and the aspect ratio of the display screen is also NTSC system 3: 4, The high-definition system has a wide-angle screen of 9:16. For this reason, a high-resolution and realistic screen can be obtained.

  Here, although it is a high vision system having such excellent characteristics, even if an NTSC system video signal is supplied as it is, an image cannot be displayed. This is because the standards differ between the NTSC system and the high vision system as described above. For this reason, when an image corresponding to an NTSC video signal is to be displayed by the high vision system, conventionally, for example, an image information conversion apparatus as shown in FIG. 9 is used to perform video signal rate conversion.

  In FIG. 9, the conventional image information conversion apparatus includes a horizontal interpolation filter 42 that performs horizontal interpolation processing on an NTSC video signal supplied via an input terminal 41, and a horizontal interpolation process. The video signal is composed of a vertical interpolation filter 43 that performs vertical interpolation processing on the video signal, and a high-definition video signal can be obtained from the output terminal 44.

  Specifically, the horizontal interpolation filter 42 has a configuration as shown in FIG. 10, and the video signal of the video system supplied via the input terminal 41 receives the first to first signals via the input terminal 51. m multipliers 52 to 52m are respectively supplied. Each multiplier 52 multiplies the video signal by coefficient data and outputs the result. The video signals multiplied by the coefficient data are supplied to first to mth adders 53 to 53m−1, respectively. Between the adders 53 to 53m-1, delay registers 54 to 54m of time T are provided, respectively. The video signal output from the mth multiplier 52m is delayed by time T by the mth delay register 54m and supplied to the m-1th adder 53m-1.

  The m-1th adder 53m-1 adds the video signal subjected to the delay time T from the mth delay register 54m and the video signal from the m-1th multiplier 52m-1. Process and output. The video signal subjected to the addition processing is again delayed by the time T by the m-1th delay register 54m-1, and the m-2th adder 53m-2 (not shown) also performs The video signal from the m-2 multiplier 53m-2 is added. In this way, the horizontal interpolation filter 42 supplies the NTSC video signal to the vertical interpolation filter 43 via the output terminal 55.

  The vertical interpolation filter 43 has the same configuration as the horizontal interpolation filter 42 described above, and performs pixel interpolation in the vertical direction on the video signal on which the horizontal interpolation processing has been performed. Accordingly, vertical pixel interpolation is performed on the NTSC video signal. The high-definition video signal thus converted is supplied to a high-definition receiver. As a result, an image corresponding to the NTSC video signal can be displayed on the high-vision receiver.

  However, since the above-described conventional image information conversion apparatus simply performs horizontal and vertical interpolation based on the NTSC video signal, the resolution is the base NTSC video signal. And nothing changed. In particular, when normal moving images are to be converted, vertical interpolation is generally performed by in-field processing, but in the case of image conversion shown in FIG. 9, there is no correlation between image fields. For this reason, the still image portion has a drawback that the resolution is deteriorated rather than the NTSC video signal due to conversion loss.

  On the other hand, the class division is performed according to the three-dimensional (spatio-temporal) distribution of the image signal level as the input signal, and the storage unit stores the prediction coefficient value acquired by learning in advance for each class, There is one that outputs an optimum estimated value by a calculation based on the calculation.

  In this method, when creating HD (High Definition) pixels, SD (Standered Definition) pixels in the same frame that are adjacent to the HD pixel to be created in the same vertical direction are divided into classes, and prediction is performed for each class. By acquiring coefficient values by learning, it is possible to obtain HD pixels closer to the true value by using intra-frame correlation in the still image portion and intra-field correlation in the motion portion. .

For example, in FIG. 11, when SD pixels x _{1 to} x ₆ are used for class classification, the quantized data resulting from performing 2-bit ADRC on the SD pixels x _{1 to} x ₆ is 3, 0, If it is 3, 0, 3, 0, it is very likely that the part is a motion image. That is, there is a high possibility that the image around the HD pixel y to be created is moving. Therefore, in the image signal conversion apparatus, the HD pixel y is created by placing importance on the SD pixel x ₂ and the SD pixel x ₄ that are close to each other in the spatially equivalent data.

On the other hand, when the quantized data obtained by performing the 2-bit ADRC on the SD pixels x _{1 to} x ₆ is 3, 3, 2, 2, 1, 0, respectively, the portion is a still image. There is a high possibility. Therefore, in the image signal conversion device, regardless of the temporal phase, the HD pixel y is created with emphasis on the SD pixel x ₂ and the SD pixel x ₄ that are close to each other in spatial position. At this time, the classification is not only with / without movement, but also expresses the signal waveform in the vicinity of the HD pixel to be created, and outputs a prediction coefficient value that is optimal for each waveform. It is.

  According to this method, the switching between still / motion can be expressed smoothly by learning using an actual image. Therefore, the occurrence of unnaturalness due to switching between still / motion as in the conventional motion adaptation method. Can be greatly reduced.

  However, the above method needs to express motion information and a waveform in space by dividing a finite number of classes. Depending on the class, patterns that should be separated are mixed in one class. There was a case.

For example, when the quantization codes obtained as a result of performing 2-bit ADRC on the above-described SD pixels x _{1 to} x ₆ are 3, 0, 3, 0, 3, 0, respectively, as described above, In the case of this pattern, the probability that the part is a motion image is extremely high. However, when the spatial frequency is rarely high, such a pattern may appear even if there is no motion. In this case, a fine horizontal stripe pattern corresponds to it. In such a case, processing for moving images is performed even for a fine horizontal stripe pattern that is stationary, that is, conversion is performed using spatially close data in a temporally equal phase, however, resolution deteriorates due to this. There was something to do.

On the other hand, the image signal conversion apparatus described in Japanese Patent Application No. 5-172617 uses three fields of data and classifies data using the values of data at the same spatial position. There is something like that. For example, as shown in FIG. 12, in addition to SD pixels x ₁ ~x _6, to classification by adding even a SD pixels x ₇ in SD pixel x ₃ spatially same position, is that.

According to this approach, it is possible to use the motion representation by data obtained by quantizing the SD pixels x ₁ and SD pixels x ₇ plus SD pixels x ₁ ~x ₆ in motion rendition by the pattern of the data obtained by quantizing. That is, in FIG. 12, when the difference value of the SD pixel x ₁ and SD pixels x ₇ is small, HD pixels y periphery of the image to create is likely to be stationary. Therefore, in the above-described image signal conversion device, the HD pixel y is created by placing emphasis on the SD pixel x ₁ and the SD pixel x ₂ that are close in spatial position. On the other hand, when the difference value between the SD pixel x ₁ and the SD pixel x ₇ is large, there is a high possibility that the image around the HD pixel y to be created is moving. Therefore, in the above-described image signal conversion apparatus, the HD pixel y is created by placing emphasis on the SD pixel x ₃ and the SD pixel x ₄ that are close in time.

  However, if there is only one set of SD pixel data in the same spatial phase, malfunctions are likely to occur due to the type of motion or the influence of noise, etc. It is desirable to use SD pixel data that is spatially in the same phase.

In addition, when there is a quick movement in the image, for example, when the movement intervenes only in one field in the image, the detection may be difficult in principle with the above-described method. That is, when the moving object passes only in the # (k + 1) field in FIG. 12, the difference value between the SD pixel x ₁ and the SD pixel x ₇ is small, and therefore, the pattern is processed closer to the stationary state. become. Therefore, in this sense, it is desirable that the sets of SD pixel data that are spatially in the same phase exist in different combinations in time.

As a countermeasure, a method of increasing the number of pixels used for class division and increasing the number of classes to reduce the quality degradation of the restored image may be considered. For example, SD pixels x ₃ spatially SD pixels x ₇ in the same position, not only, as shown in FIG. 13, SD pixels x ₈ in SD pixels x ₂ and the same spatial position, SD pixels x A method of adding an SD pixel x ₉ that is spatially the same position as ₄ can be considered. According to this method, the number of sets of SD pixel data that are spatially in the same phase increases, so that the performance of motion determination using patterns is improved.

  However, if the added 3 pixels are quantized, for example, if 2-bit ADRC is performed in the same manner as other pixels, the number of classes is 64 times that of the original method shown in FIG. As a result, the hardware scale increases and the reality is poor.

  Accordingly, an object of the present invention is made in view of the above-described problems, and generates coefficients used when converting an NTSC video signal to a high-definition video signal by improving image quality (resolution). An object of the present invention is to provide an apparatus and method for generating a coefficient.

According to a first aspect of the present invention, there is provided a coefficient generator for generating coefficient data used for converting a digital image signal having a first resolution into a digital image signal having a higher resolution and a second resolution . the data of the pixel information in the same frame of the first resolution image information in the vicinity of the pixel of interest is a product target included in resolution image information, the first image information dividing means for dividing into blocks, the first A level detection pattern is detected for each block divided by the image information dividing means, and a class detection means for outputting class detection information by determining a class belonging to each block based on the detected pattern; among the data of a plurality of frames of spatially same position of the image information of one resolution, the second image dividing the data in the vicinity of the target pixel in the block Motion that outputs motion class detection information by determining a class representing motion based on absolute value processing for difference data of a frame for each block divided by the information dividing means and the second image information dividing means Class detection means, class code generation means for determining the final class by integrating information of the class detection means and motion class detection means, reference pixels around the target pixel included in the image information of the first resolution , Coefficient calculation means for determining a coefficient for each final class by a least square method so as to minimize an error between a predicted value of the target pixel formed by linear combination with the prediction coefficient and a true value of the target pixel; It is a coefficient generator characterized by having.

According to a seventh aspect of the present invention, there is provided a coefficient generation method for generating coefficient data used for converting a digital image signal having a first resolution into a digital image signal having a higher resolution and a second resolution . A first image information dividing step for dividing the pixel information data in the same frame of the first resolution image information in the periphery of the target pixel to be generated included in the resolution image information; A level detection pattern is detected for each block divided by the image information division step, and a class detection step for outputting class detection information by determining a class belonging to each block based on the detected pattern; among the data of a plurality of frames of spatially same position of the image information of one resolution, divide the data in the vicinity of the target pixel in the block The second image information division step and the motion class detection by determining the class representing the motion based on the absolute value processing for the difference data of the frame for each block divided by the second image information division step. A motion class detection step for outputting information, a class code generation step for determining a final class by integrating information of the class detection step and the motion class detection step, and a target pixel included in the image information of the first resolution The coefficient is determined for each final class by the least square method so as to minimize the error between the predicted value of the target pixel formed by linear combination of the surrounding reference pixels and the prediction coefficient and the true value of the target pixel. And a coefficient calculation step.

  In the image information conversion apparatus according to the present invention, the input SD signal is divided by the image information dividing means into a plurality of areas composed of a plurality of pixels in the same frame continuous in the vertical direction, and the level of the image information for each area. A distribution pattern is detected, and based on the detected pattern, a class to which image information of the region belongs is determined and class detection information is output. Further, the image data is divided into a plurality of combinations of inter-frame data by different types of image information dividing means, the absolute value of the inter-frame difference at the same spatial position is calculated for each area, and a preset threshold value is calculated. The class indicating the degree of movement is determined by, and class detection information is output. The above two classes are integrated by the class code generation means and output as a final class. In the coefficient data storage means, coefficient data of a linear estimation formula, which is information for converting image information supplied from the outside into image information having a higher resolution than the image information, is stored for each class. Coefficient 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 in accordance with the coefficient data supplied from the coefficient data storage means.

  In the estimation method by classifying the spatio-temporal structure of the method proposed in the past, a failure may occur in the estimated image due to a motion mismatch. In this method, there was a drawback that the number of classes increased significantly in order to alleviate the problem, but according to the method of the present invention, the deterioration was greatly reduced while suppressing the increase in the number of classes. I can do it.

  Hereinafter, an embodiment of an image signal converter according to the present invention will be described in detail with reference to the drawings. FIG. 1 shows a schematic configuration of signal processing of this embodiment, that is, an image signal conversion apparatus. 1 is supplied as image information supplied from the outside, for example, as an SD pixel obtained by digitizing a so-called NTSC video signal.

The positional relationship between the SD pixel and the HD pixel to be created in this embodiment is shown in FIG. In other words, there are two types of HD pixels to be created: an HD pixel y ₁ that exists near the SD pixel and an HD pixel y ₂ that exists far from the SD pixel when viewed in the same field. Hereinafter, the mode for estimating the HD pixel y ₁ existing at a position close to the SD pixel is referred to as mode 1, and the mode for estimating the HD pixel y ₂ existing at a position far from the SD pixel is referred to as mode 2.

  The area dividing circuit 2 divides the SD image signal supplied from the input terminal 1 into a plurality of areas. In this embodiment, for example, three upper and lower SD pixels in the same frame of HD pixels to be created are divided into a total of 6 pixels of 1 pixel × 6 lines.

For mode 1, the SD pixels x ₁ , x ₂ , x ₃ , x ₄ , x ₅ , x ₆ corresponding to the HD pixel y ₁ in FIG. This area will be called block 1. With respect to mode 2, SD pixels x ₁ , x ₂ , x ₃ , x ₄ , x ₅ , x ₆ correspond to the HD pixel y ₂ in FIG. This area will be referred to as block 2.

  The data blocked by the area dividing circuit 2 is supplied to the ADRC circuit 3 and the delay circuit 8. The delay circuit 8 delays the data by a time required for processing of the ADRC circuit 3, the class code generation circuit 6, and the ROM table 7 and outputs the data to the estimation calculation circuit 9.

  The ADRC circuit 3 detects a one-dimensional or two-dimensional level distribution pattern of the SD data supplied for each region, and converts the data of each region into 2 bits from, for example, 8-bit SD data as described above. Pattern compression data is formed by performing operations such as compression into bit SD data, and this pattern compression data is supplied to the class code generation circuit 6.

  ADRC (Adaptive Dynamic Range Coding) is an adaptive requantization method developed for high-efficiency coding for VTR (Video Tape Recoder). In the embodiment of the present invention, it is used to generate a code for classifying a signal pattern. The ADRC circuit uses the following formula (1) as the maximum value MAX and the minimum value in the region, where DR is the dynamic range in the region, n is the bit allocation, L is the data level of the pixel in the region, and Q is the requantization code. Requantization is performed by equally dividing the value MIN with the designated bit length.

DR = MAX-MIN + 1
Q = {(L−MIN + 0.5) × 2 ⁿ / DR} (1)
However, {} means a truncation process.

In this embodiment, it is assumed that each 6-pixel SD data divided by the area dividing circuit 2 is compressed to 2 bits. Let the compressed SD data be q _{1 to} q ₆ , respectively.

  On the other hand, the SD image signal supplied from the input terminal 1 is also supplied to the area dividing circuit 4. The area dividing circuit 4 also divides the supplied SD image signal into a plurality of areas. In this embodiment, from the supplied SD image signal, for example, the upper and lower three SD pixels in the previous frame of the HD pixel to be created are divided into a total of 6 pixels of 1 pixel × 6 lines and created. For example, the upper and lower three SD pixels in the same frame as the power HD pixel are divided into a total of 6 pixels of 1 pixel × 6 lines.

That is, with respect to mode 1, the SD pixels m _{1 to} m ₆ of the previous frame and the SD pixels n _{1 to} n _{6 of the} same frame corresponding to the HD pixel y ₁ in FIG. Regarding mode 2, the SD pixels m _{1 to} m ₆ of the previous frame and the SD pixels n _{1 to} n _{6 of the} same frame corresponding to the HD pixel y ₁ in FIG.

  The data cut out by the area dividing circuit 4 is supplied to the motion class determining circuit 5. In the motion class determination circuit 5, the difference of the SD data supplied for each area is calculated, and the absolute value processing is performed on the calculated difference. An average value param of the SD data subjected to the absolute value processing is calculated, and a threshold value process is performed on the average value param, whereby a motion parameter that is a motion index is calculated, and the calculated motion parameter The motion class mv-class is determined from the above. The determined motion class mv-class is supplied to the class code generation circuit 6. Specifically, the motion class determination circuit 5 calculates the average value param of the absolute value of the difference of the supplied SD data by the following equation (2).

ただし、この実施例では、ｎ＝６である。

However, in this embodiment, n = 6.

  The motion class mv-class is determined using the average value param of the absolute value of the difference of SD data by the threshold value in which the average value param of the absolute value of the difference of the SD data calculated by the above method is set in advance. The For example, in this embodiment, four motion classes mv-class are provided, and the average value param and the motion class mv-class are set as shown in FIG.

  In the class code generation circuit 6, the following equation (3) is calculated based on the parameter compressed data supplied from the ADRC circuit 3 and the motion class mv-class supplied from the motion class determination circuit 5. The class to which the block belongs is detected, and a class code class indicating the class is supplied to the ROM table 7. This class code class indicates an address read from the ROM table 7.

この実施例では、ｎは６、Ｐは２である。

In this embodiment, n is 6 and P is 2.

In the ROM table 7, coefficient data for calculating HD data corresponding to SD data is stored for each class using a linear estimation equation by learning the relationship between the SD data pattern and HD data. Yes. This is information for converting SD data into HD data conforming to a so-called high-definition standard, which is image information having a resolution higher than that of the image information, using a linear estimation formula. In this embodiment, the coefficient data is prepared independently for mode 1 and mode 2. A method for creating coefficient data stored in the ROM table 7 will be described later. From the ROM table 7, w _i (class) that is coefficient data of the class is read from the address indicated by the class code class. The coefficient data is supplied to the estimation calculation circuit 9.

In the estimation arithmetic circuit 9, the SD data inputted from the area dividing circuit 2 and the coefficient data supplied from the ROM table 7 _wi (class) via the delay circuit 8 are inputted. HD data corresponding to the data is calculated.

More specifically, the estimation arithmetic circuit 9 uses the coefficient data for the block 1 for the mode 1 and the coefficient data for the block 1 for the mode 1 based on the SD data supplied from the delay circuit 8 and the coefficient data supplied from the ROM table 7. Calculates the HD data corresponding to the input SD data by using the coefficient data of block 2 and performing the calculation shown in the following equation (4) based on the coefficient data w _i (class). To do. The created HD data is supplied to the horizontal interpolation filter 10.

hd ′ = w ₁ x ₁ + w ₂ x ₂ + w ₃ x ₃ + w ₄ x ₄ + w ₅ x ₅ + w ₆ x ₆ (4)

  The horizontal interpolation filter 10 is the same as the horizontal interpolation filter 42 in FIG. 9, and doubles the number of pixels in the horizontal direction by interpolation processing. The output of the horizontal interpolation filter 10 is output via the output terminal 11. The HD data output via the output terminal 11 is supplied to, for example, an HD television receiver, an HD video tape recorder, or the like.

  As described above, coefficient data for estimating HD data corresponding to SD data is obtained by learning for each class in advance, and is stored in the ROM table 7 and input SD data and ROM table. The calculation is performed based on the coefficient data read out from No. 7, and HD data corresponding to the input SD data is formed and output, so that the input SD data is simply interpolated, Data closer to actual HD data can be output.

  Next, a method of creating (learning) coefficient data stored in the ROM table 7 will be described with reference to FIG. In order to obtain coefficient data by learning, first, an SD image corresponding to an already known HD image and having a ¼ pixel number of the HD image is formed. Specifically, with the ideal filter circuit shown in FIG. 8, the vertical pixel in the vertical direction of the HD data supplied via the input terminal 21 is halved by the vertical thinning filter 22. Then, the horizontal thinning filter 23 thins out pixels in the horizontal direction of the HD data to obtain SD data. The SD data obtained by the vertical thinning filter 23 is supplied to the area dividing circuit 24.

  The area dividing circuit 24 divides the SD image signal supplied from the horizontal thinning filter 23 into a plurality of areas. Specifically, the area dividing circuit 24 has the same function as the area dividing circuit 2 described above. In this embodiment, similarly to the area dividing circuit 2, the area is divided into 6 pixels. That is, the area division of the block 1 is performed for the mode 1, and the area division of the block 2 is performed for the mode 2. The SD data for each region 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 converts all or a part of the data in each area as described above. For example, an operation for compressing from 8-bit SD data to 2-bit SD data is performed to form pattern compressed data, and this pattern compressed data is supplied to the class code generation circuit 28. The ADRC circuit 25 is the same as the ADRC circuit 3 described above. In this embodiment, SD data (SD pixels x _{1 to} x ₆ in FIG. 3 and FIG. 4) of each area composed of 6 pixels divided by the area dividing circuit 24 is compressed to 2 bits by ADRC. To do.

  On the other hand, the SD pixel 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 the area dividing circuit 4 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 performs the same movement as the motion class determination circuit 5 described above. The motion class mv-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 6 described above. The pattern code data supplied from the ADRC circuit 25 and the motion class mv-class supplied from the motion class determination circuit 27 are used. Based on the calculation of the expression (2), the class to which the block belongs is detected, and the class code indicating the class is output. The class code generation circuit 28 outputs the class code to the normal equation addition 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. Hereinafter, for the sake of explanation, a case in which learning is more generalized and prediction using n pixels is performed will be described. X ₁ The SD pixel level, respectively, ‥‥, as x _n, q ₁ requantization data results of p-bit ADRC, each, ‥‥, and q _n.

At this time, the class code class of this area is defined by equation (2). Each SD pixel levels as described above, x _1, ‥‥, and x _n, when the HD pixel level was y, the coefficient data w ₁ for each class, ‥‥, the linear estimation equation of n tap according to w _n Set. This is shown in equation (5). Before learning, w _i is an undetermined coefficient.

y = w ₁ x ₁ + w ₂ x ₂ +... + w _n x _n (5)

  Learning is performed on a plurality of signal data for each class. When the number of data is m, the following equation (6) is set according to equation (5).

y _k = w ₁ x _{k 1} + w ₂ x _{k 2} +... + w _n x _kn (6)
(K = 1, 2, ... m)

When m> n, the coefficient data w _i ,..., w _n are not uniquely determined. Therefore, the coefficient data that minimizes the expression (8) by defining the element of the error vector e by the following expression (7) Ask for. This is a so-called least square method.

e _k = y _k − {w ₁ x _k1 + w ₂ x _k2 +... + w _n x _kn } (7)
(K = 1, 2, ... m)

Here, the partial differential coefficient according to w _i in equation (8) is obtained. It is only necessary to find each w _i so that the following equation (9) becomes “0”.

Hereinafter, when X _ij and Y _i are defined as in equations (10) and (11), equation (9) is rewritten into equation (12) using a matrix.

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

After the input of all learning data is completed, the normal equation adding circuit 29 outputs normal equation data to the prediction coefficient determining circuit 30. The prediction coefficient determination circuit 30 solves w _i using a general matrix solving method such as a sweeping-out method, and calculates a prediction coefficient. The prediction coefficient determination circuit 30 writes the calculated prediction coefficient in the memory 31.

As a result of learning as described above, the memory 31 stores the statistically most true value for estimating the target HD pixel y for each pattern defined by the quantized data q ₁ ,..., Q _6. A prediction coefficient that can be estimated close to is stored. As described above, the table stored in the memory 31 is the ROM table 7 used in the image signal conversion apparatus of the present invention. With the above processing, learning of coefficient data for creating HD data from SD data is completed by a linear estimation equation.

  In the description of this embodiment, the ADRC is provided as the information compression means. However, this is only an example, and any information compression means capable of expressing the signal waveform pattern in a small class is provided. For example, compression means such as DPCM (predictive coding) and VQ (vector quantization) may be used.

  Furthermore, in the description of this embodiment, for the sake of simplicity, the horizontal interpolation filter 10 is used for horizontal up-conversion. Instead, a horizontal up-conversion ROM is prepared, and horizontal up-conversion is also performed. It is of course possible to adopt the present invention using an estimation formula.

  Furthermore, in the description of this embodiment, the signal waveform pattern is expressed by one-dimensional division by the region dividing circuit 4, but it may be divided into two-dimensional divisions.

  Further, in the description of this embodiment, the motion class mv-class is determined using the SD image data that is one-dimensionally divided by the region dividing circuit 5 and the motion class determining circuit 6. The division may be a two-dimensional division. Rather, it is desirable to make it two-dimensional. In addition, for the sake of simplicity this time, the area division by the area division circuit 4 and the area division by the area division circuit 5 are similar area divisions. There is no need to do anything.

Furthermore, in the description of this embodiment, the SD pixels used for classification and the SD pixels used in the linear estimation formula are the same, but they are not necessarily the same. When different pixels are used, it is desirable that the SD pixels used in the classification are included in the SD pixels used in the linear estimation formula. In addition, the SD pixels used in the linear estimation formula used additionally are: It is desirable that only those belonging to the same field as the estimated HD pixel.

  The present invention is not limited to the above-described embodiment of the present invention, and various modifications and applications are possible without departing from the spirit of the present invention.

1 is a block diagram of an embodiment of an image information conversion apparatus according to the present invention. It is a basic diagram for demonstrating the positional relationship of SD data and HD data. It is a basic diagram for demonstrating the data used for a class division. It is a basic diagram for demonstrating the data used for a class division. It is a basic diagram for demonstrating the data used for a motion class determination. It is a basic diagram for demonstrating the data used for a motion class determination. It is a basic diagram which shows the example of the relationship between average value param and a motion class. It is an approximate line figure for explanation when creating a correction data table. It is a block diagram of the conventional image information converter. It is a circuit diagram of the principal part of the conventional image information converter. It is a basic diagram for demonstrating the problem of the conventional space-time class division system. It is a basic diagram for demonstrating the problem of the conventional space-time class division system. It is a basic diagram for demonstrating the problem of the conventional space-time class division system.

Explanation of symbols

2, 4 Area division circuit 3 ADRC circuit 5 Motion class determination circuit 6 Class code generation circuit 7 ROM table 8 Delay circuit 9 Estimation operation circuit 10 Horizontal interpolation filter

Claims (12)

In a coefficient generation device that generates coefficient data used for converting a digital image signal having a first resolution into a digital image signal having a higher resolution and a second resolution ,
First image information dividing means for dividing pixel information data in the same frame of the image information of the first resolution around the target pixel to be generated included in the image information of the second resolution into blocks;
A level distribution pattern is detected for each block divided by the first image information dividing means, and class detection information is output by determining a class belonging to each block based on the detected pattern. Class detection means;
Second image information dividing means for dividing data around the pixel of interest out of a plurality of frames of spatially the same position in the image information of the first resolution , into blocks;
A motion class detection means for outputting motion class detection information by determining a class representing motion based on absolute value processing for difference data of a frame for each block divided by the second image information dividing means; ,
Class code generating means for determining the final class by integrating information of the class detecting means and the motion class detecting means;
The error between the predicted value of the target pixel and the true value of the target pixel formed by linear combination of the reference pixels around the target pixel included in the first resolution image information and the prediction coefficient is minimized. As described above, the coefficient generating apparatus includes coefficient calculating means for determining the coefficient by the least square method for each final class.   2. The coefficient generation apparatus according to claim 1, wherein in the motion class detection means, the motion class detection information is obtained based on an average value of absolute values of difference data of the frame. The first image information dividing means has the first resolution image information in the first mode when the supplied image information and the generation target image information are close to each other. In this case, the mode is determined as the second mode , and is divided into blocks of the first mode or the second mode determined according to the position of the image information to be generated .
The class detection means outputs class detection information of each block in the first and second modes from the first image information dividing means,
The second image information dividing unit determines the image information of the first resolution from a range larger than the first image information dividing unit according to the position of the image information to be generated. Split into modes or blocks of the second mode,
The motion class detection means outputs the motion class detection information of each block in the first and second modes from the second image information dividing means,
The class code generating means determines the final class by integrating the information of the class detecting means and the motion class detecting means in each of the first and second modes ,
2. The coefficient generation apparatus according to claim 1, wherein the coefficient calculation means determines the coefficient for each final class in each of the first and second modes . The coefficient generation apparatus according to claim 3 , wherein the large range is a large range in a space-time direction of the supplied image information.   2. The coefficient generation apparatus according to claim 1, further comprising coefficient data storage means for storing the generated coefficient for each mode.   The coefficient generation apparatus according to claim 1, wherein the coefficient data is generated for each mode. In a coefficient generation method for generating coefficient data used for converting a digital image signal having a first resolution into a digital image signal having a higher resolution and a second resolution ,
A first image information dividing step of dividing pixel information data in the same frame of the image information of the first resolution around the target pixel to be generated included in the image information of the second resolution into blocks;
A level distribution pattern is detected for each block divided by the first image information dividing step, and class detection information is output by determining a class belonging to each block based on the detected pattern. A class detection step;
A second image information dividing step of dividing the data around the pixel of interest out of a plurality of frames of spatially the same position of the image information of the first resolution into blocks;
A motion class detection step for outputting motion class detection information by determining a class representing motion based on absolute value processing for difference data of a frame for each block divided by the second image information division step; ,
A class code generation step for determining a final class by integrating information of the class detection step and the motion class detection step;
The error between the predicted value of the target pixel and the true value of the target pixel formed by linear combination of the reference pixels around the target pixel included in the first resolution image information and the prediction coefficient is minimized. As described above, the coefficient generation method includes a coefficient calculation step for determining the coefficient by the least square method for each final class. 8. The coefficient generation method according to claim 7 , wherein, in the motion class detection step, the motion class detection information is obtained based on an average value of absolute values of difference data of the frame. In the first image information dividing step, the image information of the first resolution exists in a first mode when the supplied image information and the image information to be generated are close to each other, and in a distant position. In this case, the mode is determined as the second mode , and is divided into blocks of the first mode or the second mode determined according to the position of the image information to be generated .
The class detection step outputs class detection information of each block in the first and second modes from the first image information division step.
In the second image information dividing step, the first resolution image information is determined in accordance with the position of the image information to be generated from a range larger than the first image information dividing step. Split into modes or blocks of the second mode,
The motion class detection step outputs the motion class detection information of each block in the first and second modes from the second image information division step,
The class code generation step determines the final class by integrating the information of the class detection step and the motion class detection step in each of the first and second modes .
8. The coefficient generation method according to claim 7 , wherein in the coefficient calculation step, the coefficient is determined for each final class in each of the first and second modes . The coefficient generation method according to claim 9 , wherein the large range is a large range in a space-time direction of the supplied image information. 8. The coefficient generation method according to claim 7 , further comprising a coefficient data storage step, wherein the generated coefficient is stored for each mode. The coefficient generation method according to claim 7 , wherein the coefficient data is generated for each mode.

Images