JP2000152273A - Image information converter and converting method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enhance the image quality in the case of converting a composite signal into a component signal through the use of classification adaptation processing. SOLUTION: An NTSC signal being a composite signal is fed to area extract circuits 51, 55, which extract a class tap and a prediction tap respectively. A pattern detection circuit detection circuit 52 conducts an arithmetic in response to a phase of pixels such as obtaining a mean value of pixel values of pixels with inverted phases to each other based on an output of the area extract circuit 51 and detects a pattern of the NTSC signal based on the arithmetic result. A class code decision circuit 53 generates a class code based on the detection result. A coefficient memory 54 outputs that corresponding to the class code among prediction coefficients for each class stored in advance. A prediction arithmetic section 56 uses the prediction tap extracted by the area extract circuit 55 and the output of the coefficient memory 54 to conduct prediction arithmetic operation and generates, e.g. a luminance signal Y as a component signal.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image information conversion apparatus and a conversion method which can be used for conversion for generating a component signal from a composite signal.

[0002]

2. Description of the Related Art For example, a television signal of the NTSC system is obtained by multiplexing a luminance signal Y and a chroma signal C with balanced modulation. For this reason, a conventional television receiver receives a television signal and receives a CRT (Cathod).
When displaying a color image on a display unit such as (Ray Tube),
By performing Y / C separation for separating the luminance signal Y and the chroma signal C from the television signal and further performing a process of demodulating the chroma signal C, the luminance signal Y and the color difference signals RY and BY are obtained. By obtaining the component signal of
It is necessary to generate a GB signal. Therefore, there is a problem that the scale of the processing circuit becomes large.

In a conventional Y / C separation method, a filtering process using a two-dimensional Y / C separation circuit, a three-dimensional Y / C separation circuit, or the like, particularly in an edge portion of an image or a moving image portion, There is a problem that image quality is likely to deteriorate due to Y / C separation errors such as dot disturbance and cross color.

In order to improve the above problems,
It is conceivable to convert a composite signal such as an NTSC signal into a luminance signal Y, color difference signals RY, BY, and the like by using a classification adaptive process. In this case, first, class classification is performed using features obtained from pixels having a predetermined positional relationship with the target pixel in the composite signal. Then, a component signal is directly obtained from the composite signal by an operation based on a coefficient prepared in advance for each class to be classified and a pixel having a predetermined positional relationship with the pixel of interest.

[0005]

In such processing, when an appropriate coefficient cannot be obtained for the pixel of interest due to the nature of the composite signal, dot interference,
There is a problem that image quality deterioration such as cross color cannot be sufficiently improved.

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide an image information conversion apparatus and a conversion method capable of further reducing image quality deterioration when converting a composite signal into a component signal using the class classification adaptive processing. is there.

[0007]

According to a first aspect of the present invention, there is provided an image information conversion apparatus for performing image information conversion on an input image signal, wherein a target pixel and a target pixel have a predetermined positional relationship from the input image signal. A first image data selecting means for selecting a pixel, and an output from the first image data selecting means,
A pattern detection unit that performs a predetermined calculation process related to the phase of the predetermined pixel and detects a pattern in the vicinity of the pixel of interest based on a result of the calculation process; Class code generating means for generating a class code indicating a class; second image data selecting means for selecting a pixel having a predetermined positional relationship with respect to a pixel of interest from an input image signal; A coefficient storage means for storing a prediction coefficient and outputting a prediction coefficient corresponding to an output of the class code generation means; a predetermined arithmetic processing based on an output of the coefficient storage means and an output of the second image data selection means And a prediction calculation processing means for performing the following.

According to a tenth aspect of the present invention, in the image information conversion method for performing image information conversion on an input image signal, a first pixel and a pixel having a predetermined positional relationship with respect to the target pixel are selected from the input image signal. Image data selecting step of
A pattern detection step of performing a predetermined calculation process relating to a phase of a predetermined pixel based on a result of the first image data selection step, and detecting a pattern near a pixel of interest based on a result of the calculation process; A class code generating step of generating a class code indicating a class in the vicinity of the pixel of interest based on a result of the step; and a second image for selecting a pixel having a predetermined positional relationship with respect to the pixel of interest from the input image signal A data selection step, a coefficient storage step of storing a prediction coefficient predetermined for each class, and outputting a prediction coefficient corresponding to a result of the class code generation step, a result of the coefficient storage step, and a second image data selection step. A prediction operation processing step of performing a predetermined operation processing based on the result of the step. It is a method.

According to the invention described above, even when a composite image signal in which each pixel has a different phase depending on a position, such as an NTSC signal, is used as an input image signal, the result of an operation based on pixels near the pixel of interest is obtained. Can be used to make the appropriate classification.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Prior to the description of an embodiment of the present invention, in order to facilitate understanding, the general configuration of a general NTSC television receiver will be described with reference to FIG. explain. The receiving antenna 11 captures the radio wave and supplies a signal based on the captured radio wave to the tuner 12. The tuner 12 selects and amplifies a desired channel from the supplied signals, and generates an intermediate frequency signal. The intermediate frequency signal is supplied to the intermediate frequency signal processing circuit 13.
The intermediate frequency signal processing circuit 13 demodulates and supplies the supplied signal.
The composite signal is detected to generate a composite signal whose gain is appropriately adjusted.
To supply.

The Y / C separation circuit 14 separates the supplied composite signal into a luminance signal Y and a chrominance signal C, supplies the chrominance signal C to a chroma demodulation circuit 15, and supplies the luminance signal Y to a matrix circuit 16. I do. The chroma demodulation circuit 15 performs color demodulation of the color signal C to generate baseband RY and BY signals, and supplies the generated RY and BY signals to the matrix circuit 16. The matrix circuit 16 performs matrix processing on the supplied signals to generate primary color R, G, and B signals. Then, the generated R, G, B signals are converted to CR signals.
Supply to T17.

The configuration shown in FIG. 1 is such that the NTSC signal as a composite signal is processed by the Y / C separation circuit 14, the chroma demodulation circuit 15 and the matrix circuit 16 to obtain the primary color RGB signals. However, there is a problem that the circuit scale becomes large. In addition, there is a problem that image quality is likely to be deteriorated due to an operation error of the Y / C separation circuit 14 such as dot disturbance or cross color at an edge of an image or a moving image portion.

According to the present invention, a processing system 18 including a Y / C separation circuit 14 and a chroma demodulation circuit 15 in FIG. 1 or a Y / C separation circuit 14, a chroma demodulation circuit 15, The processing performed by the processing system 19 including the circuit 16 is collectively performed.

FIG. 2 shows an example of the overall configuration according to an embodiment of the present invention. Receiving antenna 11, a tuner 12 and intermediate frequency composite signal generated by the processing sequence is performed by the amplifier circuit 13 is supplied to the A / D converter 21, for example, the color subcarrier frequency f 4 times the sampling frequency of the _SC (4f _SC ), And is supplied to the classification adaptive processing unit 22. The classification adaptive processing unit 22 includes baseband Y, RY, BY
A component signal is generated, and the generated Y, RY, B
-Supply the Y component signal to the matrix circuit 23; The matrix circuit 23 generates primary color signals R, G, and B based on the supplied signals, and supplies the generated primary color signals R, G, and B to the CRT 17.

FIG. 3 shows another example of the overall configuration according to the embodiment of the present invention. A signal generated by processing sequentially performed by the receiving antenna 11, the tuner 12, and the intermediate frequency amplification circuit 13 is converted into an A / D converter 3.
1, A / D converted, and supplied to the classification adaptive processing unit 32. The classification adaptive processing unit 32 generates primary color signals R, G, and B based on the supplied signals,
The generated primary color signals R, G, B are supplied to the CRT 17.

The configuration of the classification adaptive processing unit 22 or the classification adaptive processing unit 32 will be described with reference to FIG. The NTSC signal as a composite signal is supplied to the area extraction circuits 51 and 55. The region extraction circuit 51
A class tap, that is, a pixel necessary for performing class classification is extracted from the supplied NTSC signal, and the class tap is supplied to the pattern detection circuit 52. The pattern detection circuit 52 performs an arithmetic process as described below with respect to the phase of each pixel supplied as a class tap, detects a pattern of an NTSC signal based on the result of the arithmetic process, and classifies the detection result with a class code. It is supplied to the decision circuit 53.

The class code determination circuit 53 determines a class based on the detection result supplied from the pattern detection circuit 52, and generates a class code corresponding to the determined class. The generated class code is supplied to the coefficient memory 54. The coefficient memory 54 stores in advance prediction coefficients for each class obtained by learning described later, and outputs a prediction coefficient corresponding to the supplied class code from the stored contents. The output of the coefficient memory 54 is the prediction operation circuit 5
6.

On the other hand, the area extraction circuit 55
A pixel having a predetermined positional relationship with the target pixel is extracted from the TSC signal, and pixel data relating to the extracted pixel is supplied to the prediction calculation circuit 56. The prediction calculation circuit 56 generates a luminance signal Y by performing a prediction calculation according to the following equation (1) using the supplied pixel data and the output of the coefficient memory 54. The pixels used for the prediction calculation (pixels extracted by the region extraction circuit 55) are called prediction taps.

Y = w ₁ × x ₁ + w ₂ × x ₂ + ‥‥ + w _n × x _n (1) Here, y is a target pixel to be predicted and estimated. Also,
x ₁ , ‥‥, x _n are each prediction tap, w ₁ , ‥‥,
w _n is each prediction coefficient.

Next, FIG. 5 shows an example of a configuration for learning, that is, obtaining a prediction coefficient for each class. Generally, in order to perform learning, a signal having the same signal format as a signal to be generated by the classification adaptive processing (hereinafter, referred to as a teacher signal) is input, and generated based on the teacher signal and the teacher signal. The prediction coefficient is determined by an operation based on a signal to be input in the classification adaptive processing and a signal having the same signal format (hereinafter, referred to as a student signal).

Here, a luminance signal Y is input as a teacher signal. The luminance signal Y is supplied to the NTSC encoder 61 and the normal equation generation circuit 66. Further, the color difference signals RY and BY are supplied to the NTSC encoder 61. The NTSC encoder 61 receives the supplied Y, R-
Y, BY are converted into NTSC signals as student signals,
This NTSC signal is supplied to region extraction circuits 62 and 65. The area extraction circuit 62 extracts a class tap from the supplied NTSC signal, and supplies the extracted class tap to the pattern detection circuit 63. The pattern detection circuit 63
A pattern of the NTSC signal is detected based on the supplied class tap, and the detection result is used as a class code determination circuit 64.
To supply.

The class code determination circuit 64 determines a class based on the detection result supplied from the pattern detection circuit 63, and generates a class code corresponding to the determined class. The generated class code is transmitted to the normal equation generation unit 66.
Supplied to On the other hand, the region extraction circuit 65 extracts prediction taps from the supplied NTSC signal, and supplies the extracted prediction taps to the normal equation generation unit 66. The normal equation generation unit 66 generates data necessary for solving the normal equation based on the luminance signal Y and the prediction tap supplied from the area extraction circuit 65. The normal equation is an equation using a prediction coefficient for each class as a solution, as described later.

The output of the normal equation generator 66 is supplied to a coefficient determining circuit 67. When sufficient data is supplied, the coefficient determination circuit 67 solves the normal equation using, for example, the least squares method, thereby obtaining the prediction coefficients w ₁ , ‥‥, w _n
Is calculated. The calculated prediction coefficients are supplied to the memory 68 and stored. The prediction coefficients stored in the memory 68 are loaded into the coefficient memory 54 in FIG. Note that the above description is for the case of generating a prediction coefficient for predicting the luminance signal Y, but RY, BY, R, G,
In order to generate a prediction coefficient for predicting B or the like, a signal to be predicted may be input as a teacher signal instead of Y.

The normal equation will be described. In the above equation (1), before learning, the prediction coefficients w ₁ , ‥‥, and w _n are undetermined coefficients. Learning is performed by inputting a plurality of teacher signals for each class. When m types are input as the teacher signal, the following equation (2) is set from the equation (1).

Y _k = w ₁ × x _k1 + w ₂ × x _k2 + ‥‥ + w _n × x _kn (2) (k = 1,2, ‥‥, m) When m> n, the prediction coefficients w ₁ , Since ‥‥ and w _n are not uniquely determined, the element e _k of the error vector e is defined by the following equation (3), and the prediction coefficient is set so as to minimize the error vector e defined by the equation (4). To be determined. That is, the prediction coefficient is uniquely determined by the so-called least square method.

E _k = y _k − {w ₁ × x _k1 + w ₂ × x _k2 + Δ + w _n × x _kn } (3) (k = 1, 2, Δm)

[0027]

(Equation 1)

As a practical calculation method for obtaining the prediction coefficient minimizing e ² in Equation (4), partial differentiation of e ² with the prediction coefficient w _i (i = 1, 2 ‥‥) is performed (Equation 4). (5)) Each prediction coefficient w _i may be determined so that the partial differential value becomes 0 for each value of _i .

[0029]

(Equation 2)

A specific procedure for determining each prediction coefficient w _i from equation (5) will be described. If X _ji and Y _i are defined as in equations (6) and (7), equation (5) can be written in the form of a determinant of equation (8).

[0031]

(Equation 3)

[0032]

(Equation 4)

[0033]

(Equation 5)

Equation (8) is generally called a normal equation. Here, X _ji (j, i = 1, 2 ‥‥ n) and Y _i (i = 1, 2 ‥‥ n) are calculated based on the teacher signal and the student signal. That is, the normal equation generation circuit 66 calculates the values of X _ji and Y _i to determine the normal equation (8), and the coefficient determination circuit 67 solves the normal equation (8) to determine each prediction coefficient w _i . Perform processing. Note that the prediction coefficient is generated for each chroma phase in consideration of the chroma phase (four types of sampling phases under a sampling frequency of 4f _SC ) in each pixel in the NTSC signal as a composite signal.

In the above-described class classification adaptive processing, as a method of obtaining the characteristics of the waveform of an input image signal in order to perform class classification, the pattern detection circuit 52 uses, for example, an ADRC (Adaptive Dynamic Range Coding).
A) A method of requantizing the image data of the pixel extracted as a class tap by performing the processing is used. In particular, when class classification adaptive processing is used to generate a component signal from a composite signal such as an NTSC signal, Y / C signals such as dot disturbance, cross color, etc.
Keeping in mind that the image quality degradation due to the separation error occurs at the edge portion (change point) of the image, it is considered effective to perform class classification that captures the change point.

However, in the NTSC signal, since the chroma signal C is balanced-modulated and multiplexed on the luminance signal Y,
Even if ADRC processing is simply performed on the NTSC signal,
Since the color phase between pixels differs due to the influence of the color subcarrier signal, accurate waveform classification cannot be performed. In order to perform effective classification in consideration of such a problem, ADR is performed using only pixels having the same color phase.
There has been proposed a method of performing class classification by C processing or the like. However, with such a method, it is difficult to see a pattern near the pixel of interest.

This will be described with reference to FIG.
FIG. 6 is a schematic diagram showing the phase of the chroma signal C when the NTSC signal is sampled at a sampling frequency (4f _SC ) four times the color subcarrier frequency f _SC . FIG. 6A illustrates the field two times before the present, and FIG. 6B illustrates the current field. Here, the vertical direction and the horizontal direction correspond to the vertical and horizontal directions in the image, respectively, and a black circle, a black square, a white circle,
Each white square is 0 with respect to the chroma phase of the pixel of interest.
Degrees, +90 degrees, +180 degrees, and +270 degrees.
In FIGS. 6A and 6B, the phase is shifted by 180 degrees between the pixels at the corresponding positions. As can be seen from FIG. 6B, the pixel having the same phase as the target pixel is not included in the eight pixels adjacent to the target pixel, and the nearest pixel is the second pixel counted from the target pixel in the vertical direction. Things.

Therefore, according to the present invention, the pattern detection circuit 52 performs simple Y / C separation using pixels that are temporally or spatially close to the target pixel, and performs ADRC processing based on the result of the simple Y / C separation. And so on to perform class classification. The simple Y / C separation is an operation process as described below. First, an average value of pixel values of pixels having mutually opposite phases with respect to color is regarded as a luminance signal, and Y
And Further, a value obtained by subtracting the pixel value of the pixel of the opposite phase from the pixel value of the pixel of the same phase with respect to the color of the target pixel and dividing by 2 is regarded as a color difference signal, and is represented by C. Here, the color difference signal is
Unlike the luminance signal, when the average of the pixel values of the pixels having the opposite phases with respect to the color is taken, the influence of the subcarrier of the color cannot be removed. Therefore, C is calculated as described above.

For example, as shown in FIG. 7, a pixel of interest indicated by two oblique lines crossing each other is a ₅ and two fields before a ₄ and a ₆ on the same line as a _5. the results of the simplified Y / C separation and Y _1, Y _2, Y ₃ and C _1, C _2, C ₃ with the pixels at the same position (i.e. _{a 11, a 10, a 12} ). Where Y _i and C _i (i
= 1, 2, 3) is obtained by the following equation.

Y ₁ = (a ₄ + a ₁₀ ) / 2, Y ₂ = (a ₅ + a ₁₁ ) / 2, Y ₃ = (a ₆ + a ₁₂ ) / 2 (9) C ₁ = (a _{4 - a 10) / 2, C} 2 = (a 5 - a 11) / 2, C 3 = (a 6 - a 12) / 2 (10) Moreover, the same line on the a ₅ and a ₅ in a similar manner a _4, 1 line before the pixels at the same position relative to a ₆ in (a
₂ , a ₁ , a ₃ ), the results of simple Y / C separation are defined as Y ₄ , Y ₅ , Y ₆ and C ₄ , C ₅ , C ₆ . Here, Y _i and C _i (i = 4, 5, 6) are obtained by the following equations.

Y ₄ = (a ₄ + a ₁ ) / 2, Y ₂ = (a ₅ + a ₂ ) / 2, Y ₃ = (a ₆ + a ₃ ) / 2 (11) C ₄ = (a ₁ -a ₄ ) / 2, C ₂ = (a ₅ -a ₂ ) / 2, C ₃ = (a ₃ -a ₆ ) / 2 (12) In the same way, on the same line as a ₅ and a ₅ At the same position after one line with respect to a ₄ and a ₆ (a
_8, a _7, a ₉₎ The results of the simplified Y / C separation and Y _7, Y _8, Y ₉ and C _7, C _8, C ₉ with. Here, Y _i and C _i (i = 7, 8, 9) are obtained by the following equations.

Y ₇ = (a ₄ + a ₇ ) / 2, Y ₈ = (a ₅ + a ₈ ) / 2, Y ₉ = (a ₆ + a ₉ ) / 2 (13) C ₇ = (a _{7 - a 4) / 2, C} 2 = (a 5 - a 8) / 2, C 3 = (a 9 - a 6) / 2 (14) thus obtained, the simplified Y / C separation output Y
_i (i = 1, 2, ‥‥, 9) and C _i (i = 1, 2, 2)
{, 9) include the characteristic of the waveform in the vicinity of the pixel of interest in the area surrounded by the dotted line in FIG. These Y _i or C _i, obtains the features in the waveform using the ADRC, performing classification.

An example of the actual class classification will be described below. When the image in the area surrounded by the dotted line in FIG. 7 is uniform, the simplified Y / C separation output Y _i is also uniform. Conversely, when there is a change in the image in the area surrounded by the dotted line in FIG. 7, Y _i is a value reflecting the change. Therefore, YRC is performed by using an ADRC process as shown below.
By performing the class classification according to the characteristic of _i, the characteristic of the waveform near the target pixel can be reflected in the classification result.

The _{code i = (Y i -MIN)} × 2 n / DR DR = MAX-MIN + 1 where, code _i is ADRC code, MIN is the minimum value of Y _i, MAX is the maximum value of Y _i, DR is Y _{i represents} the dynamic range, and n represents the number of requantization bits.

For example, as shown in FIG.
‥‥, 9, when n = 1 (ie, 1
For bit ADRC), it is classification into one of 2 ⁹ patterns based on the value of Y _i. As for the C _i, as in the case of Y _i, C _i using ADRC processing
By performing the class classification in accordance with the characteristics described above, the characteristics of the waveform in the vicinity of the target pixel can be reflected in the classification result.

An example of a method of generating a component signal from an NTSC signal as a composite signal by performing the class classification adaptive processing based on the simplified Y / C separation output as described above will be described with reference to the flowchart of FIG. explain. At step S1, a target pixel is selected in the NTSC signal. In step S2, the pixel of interest and its neighboring pixels are extracted. In step S3, the simplified Y / Y is calculated by the calculation based on the above-described equations (9) to (14) based on the pixel data extracted in step S2.
C separation is performed. Then, as step S4, simple Y /
An ADRC process is performed on the C separation output to generate an ADRC code, and further, as step S5, class classification is performed based on the ADRC code.

Next, as step S6, step S5
The prediction coefficient corresponding to the result of the class classification by is read out. Here, the prediction coefficient to be read is stored in a predetermined storage circuit such as the coefficient memory 54 in FIG. 4 in advance. On the other hand, as step S7, a prediction tap is extracted from the pixel of interest and its neighboring pixels in the NTSC signal. In step S8, a prediction operation according to the above-described equation (1) or the like is performed using the prediction coefficient read in step S6 and the prediction tap extracted in step S7.

Further, as step S9, step S
8 to output a component signal such as a luminance signal Y predicted and generated. Then, as step S10, it is determined whether or not the processing of steps S1 to S9 has been completed for all the pixels. If it is determined that the processing has been completed for all the pixels, the processing is ended. Otherwise, the process proceeds to step S1, and the processing is performed for the unprocessed pixels.

FIG. 9 shows an example of a tap structure of a prediction tap according to an embodiment of the present invention. Here, the target pixel is shown with two oblique lines crossing each other. Here, a total of 20 taps (shown by a dotted line) of 15 pixels in the current field including the pixel of interest and 5 pixels in the field two fields before, and an offset component A total of 21 pixels of one assigned tap are used as prediction taps. Here, the offset component is a DC-like shift generated between the composite signal and the component signal in the process of generating the component signal from the composite signal.

The prediction tap assigned to the offset component can be handled in the same manner as other prediction taps. Further, since such an offset component can be obtained by theoretical calculation, a theoretical value may be fixedly used as data of a prediction tap assigned to the offset component. Generally, the same operation accuracy can be obtained in practice by using any of the two treatments described above for the prediction tap.

The above description is for the case where the luminance signal Y as a component signal is predicted from the NTSC signal as a composite signal. However, the present invention is also applicable to a case where a signal other than the NTSC signal is used as a composite signal. It is possible to apply. The present invention can also be applied to the case where RY, BY, R, G, B, etc. are predicted as component signals.
However, in this case, it is necessary to perform learning for generating a prediction coefficient corresponding to the component signal to be predicted. Further, the color signal C may be obtained by subtracting the predicted luminance signal Y from the original composite signal, or conversely, the color signal C is predicted by the class classification adaptive processing, and the predicted color signal C is converted to the original composite signal. The color signal C may be obtained by subtracting from the signal.

The above-described embodiment of the present invention
Although the present invention is applied to a case where image information conversion for generating a component signal from a composite signal is performed in a television receiver, the application of the present invention is not limited to this. For example,
The present invention can be applied to a case where image information conversion is performed in a tuner or adapter for a television receiver, a video recorder such as a VTR, and equipment for broadcasting.

[0053]

As described above, according to the present invention, when performing image information conversion such as generation of a component signal from a composite signal such as an NTSC signal using the classification adaptive processing, pixel data extracted as a class tap is obtained. In addition, a predetermined arithmetic process related to the phase of each pixel is performed, and a 1-bit ADRC process or the like is performed based on the result of the arithmetic process to perform class classification.

Therefore, even when an image signal in which each pixel has a different phase depending on a position, such as an NTSC signal, is used as an input image signal, an appropriate result can be obtained by using a calculation result based on a pixel near the target pixel. Classification can be performed.

For this reason, the circuit scale can be reduced and the Y / C ratio can be reduced as compared with a configuration in which Y / C separation is performed on the composite signal and a component signal is obtained by matrix processing or the like thereafter. Various deteriorations such as dot disturbance and cross color caused by the separation error can be reduced.

In particular, pixel data at a position closer to the target pixel can be extracted as a class tap as compared with a case where a pixel having the same phase as the target pixel is used as a class tap. It can be done more accurately. For this reason, it is possible to more appropriately reduce the image quality deterioration.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an example of the overall configuration of a general NTSC television receiver.

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

FIG. 3 is a block diagram showing another example of the overall configuration according to the embodiment of the present invention.

FIG. 4 is a block diagram illustrating an example of a configuration related to prediction estimation according to an embodiment of the present invention.

FIG. 5 is a block diagram illustrating an example of a configuration related to learning according to an embodiment of the present invention.

FIG. 6 is a schematic diagram for explaining a phase of a pixel in an NTSC signal.

FIG. 7 is a schematic diagram illustrating an example of pixel positions extracted for performing class classification in one embodiment of the present invention.

FIG. 8 is a flowchart illustrating a process according to an embodiment of the present invention.

FIG. 9 is a schematic diagram illustrating an example of a tap structure of a prediction tap according to the embodiment of the present invention.

[Explanation of symbols]

22, 32 ... classification adaptive processing unit, 52 ... pattern detection circuit

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] January 29, 1999 (1999.1.2
9)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] 0026

[Correction method] Change

[Correction contents]

E _k = y _k − {w ₁ × x _k1 + w ₂ × x _k2 + ‥‥ + w _n × x _kn } (3) (k = 1, 2, ‥‥ , m)

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0031

[Correction method] Change

[Correction contents]

[0031]

(Equation 3)

[Procedure amendment 3]

[Document name to be amended] Statement

[Correction target item name] 0032

[Correction method] Change

[Correction contents]

[0032]

(Equation 4)

[Procedure amendment 4]

[Document name to be amended] Statement

[Correction target item name] 0034

[Correction method] Change

[Correction contents]

Equation (8) is generally called a normal equation. Here, X _ji (j, i = 1 , 2 ‥‥ , n),
And Y _i (i = 1,2 , n) are calculated based on the teacher signal and the student signal. That is, the normal equation generation circuit 66 calculates the values of X _ji and Y _i to determine the normal equation (8), and the coefficient determination circuit 67 solves the normal equation (8) to determine each prediction coefficient w _i . Perform processing. Note that the prediction coefficient is generated for each chroma phase in consideration of the chroma phase (four types of sampling phases under a sampling frequency of 4f _SC ) in each pixel in the NTSC signal as a composite signal.

[Procedure amendment 5]

[Document name to be amended] Statement

[Correction target item name] 0038

[Correction method] Change

[Correction contents]

Therefore, according to the present invention, the pattern detection circuit 52 performs simple Y / C separation using pixels that are temporally or spatially close to the target pixel, and performs ADRC processing based on the result of the simple Y / C separation. And so on to perform class classification. The simple Y / C separation is an operation process as described below. First, the average value of pixel values of pixels having mutually opposite phases with respect to color is regarded as a luminance signal, and Y
And Further, a value obtained by subtracting the pixel value of the pixel of the opposite phase from the pixel value of the pixel of the same phase with respect to the color of the target pixel and dividing by 2 is regarded as a color difference signal, and is represented by C. Calculate C in this way
Reduces the effects of color subcarriers
A difference signal equivalent can be obtained.

[Procedure amendment 6]

[Document name to be amended] Statement

[Correction target item name] 0039

[Correction method] Change

[Correction contents]

For example, as shown in FIG. 7, a target pixel a ₅ indicated by two oblique lines intersecting each other, and a ₄ and a ₆ on the same line as a ₅ two fields before. The result of performing simple Y / C separation using pixels at the same position (ie, a ₁₁ , a ₁₀ , a ₁₂ ) is represented by Y ₁ , Y ₂ , Y ₃ and C
₁ , C ₂ and C ₃ . Here, Y _i and C _i (i =
1, 2, 3) is obtained by the following equation.

[Procedure amendment 7]

[Document name to be amended] Statement

[Correction target item name] 0040

[Correction method] Change

[Correction contents]

Y ₁ = (a ₄ + a ₁₀ ) / 2, Y ₂ = (a ₅ + a ₁₁ ) / 2, Y ₃ = (a ₆ + a ₁₂ ) / 2 (9) C ₁ = ( a ₁₀ -a ₄ ) / 2, C ₂ = (a ₅ -a ₁₁ ) / 2, C ₃ = ( a ₁₂ -a ₆ ) / 2 (10) In the same manner, on the same line as a ₅ and a ₅ a _4, 1 line before the pixels at the same position relative to a ₆ in (a
₂ , a ₁ , a ₃ ), the results of simple Y / C separation are defined as Y ₄ , Y ₅ , Y ₆ and C ₄ , C ₅ , C ₆ . Here, Y _i and C _i (i = 4, 5, 6) are obtained by the following equations.

[Procedure amendment 8]

[Document name to be amended] Statement

[Correction target item name] 0041

[Correction method] Change

[Correction contents]

Y ₄ = (a ₄ + a ₁ ) / 2, Y ₅ = (a ₅ + a ₂ ) / 2, Y ₆ = (a ₆ + a ₃ ) / 2 (11) C ₄ = (a _{1 - a 4) / 2, C} 5 = (a 5 - a 2) / 2, C 6 = (A ₃ -a ₆ ) / 2 (12) Further, in the same manner, pixels (a) at the same position after one line with respect to a ₄ and a ₆ on the same line as a ₅ and a _5.
8, a _7, a ₉₎ The results of the simplified Y / C separation and Y _7, Y _8, Y ₉ and C _7, C _8, C ₉ with. Here, Y _i and C _i (i = 7, 8, 9) are obtained by the following equations.

[Procedure amendment 9]

[Document name to be amended] Statement

[Correction target item name] 0042

[Correction method] Change

[Correction contents]

Y ₇ = (a ₄ + a ₇ ) / 2, Y ₈ = (a ₅ + a ₈ ) / 2, Y ₉ = (a ₆ + a ₉ ) / 2 (13) C ₇ = (a _{7 - a 4) / 2, C} 8 = (a 5 - a 8) / 2, C 9 = (A ₉ -a ₆ ) / 2 (14) Y of the simplified Y / C separation output obtained in this way
_i (i = 1, 2, ‥‥, 9) and C _i (i = 1, 2, 2)
{, 9) include the characteristic of the waveform in the vicinity of the pixel of interest in the area surrounded by the dotted line in FIG. These Y _i or C _i, obtains the characteristics of the waveform using the ADRC, performing classification.

[Procedure amendment 10]

[Document name to be amended] Statement

[Correction target item name] 0051

[Correction method] Change

[Correction contents]

The above description is for the case where the luminance signal Y as a component signal is predicted from the NTSC signal as a composite signal. However, the present invention is also applicable to a case where a signal other than the NTSC signal is used as a composite signal. It is possible to apply. The present invention can also be applied to the case where RY, BY, R, G, B, etc. are predicted as component signals.
However, in this case, it is necessary to perform learning for generating a prediction coefficient corresponding to the component signal to be predicted. Further, the color signal C may be obtained by subtracting the predicted luminance signal Y from the original composite signal, or conversely, the color signal C is predicted by the class classification adaptive processing, and the predicted color signal C is converted to the original composite signal. The luminance signal Y may be obtained by subtracting from the signal.

[Procedure amendment 11]

[Document name to be amended] Drawing

[Correction target item name] Figure 2

[Correction method] Change

[Correction contents]

FIG. 2

[Procedure amendment 12]

[Document name to be amended] Drawing

[Correction target item name] Figure 3

[Correction method] Change

[Correction contents]

FIG. 3

[Procedure amendment 13]

[Document name to be amended] Drawing

[Correction target item name] Fig. 7

[Correction method] Change

[Correction contents]

FIG. 7

Continuation of the front page (72) Inventor Hideo Nakaya 6-7-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Inside Sony Corporation (72) Inventor Takaya Hoshino 6-35, Kita-Shinagawa, Shinagawa-ku, Tokyo Sony (72) Inventor Ken Sou Inoue 6-7-35 Kita Shinagawa, Shinagawa-ku, Tokyo Sony Corporation (72) Inventor Takeharu Nishikata 6-35, Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation (72) Inventor Naoki Kobayashi 6-7-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation (72) Inventor Tetsushi Kokubo 6-35-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation F Term (reference) 5C066 AA03 BA02 CA05 DC01 DC06 GA02 GA05 HA02 KE01 KE07 KE19

Claims (10)

[Claims] 1. An image information conversion apparatus for performing image information conversion on an input image signal, comprising: a first image data selection unit for selecting a target pixel and a pixel having a predetermined positional relationship with respect to the target pixel from the input image signal; Means for performing a predetermined arithmetic processing relating to the phase of the predetermined pixel based on an output of the first image data selecting means, and detecting a pattern near the pixel of interest based on a result of the arithmetic processing Pattern detection means; class code generation means for generating a class code indicating a class near the pixel of interest based on the output of the pattern detection means; and a predetermined position with respect to the pixel of interest from the input image signal. A second image data selecting means for selecting a pixel having a relationship, and a prediction coefficient predetermined for each of the classes are stored. A coefficient storage unit for outputting a prediction coefficient corresponding to the following, and a prediction calculation processing unit for performing a predetermined calculation process based on an output of the coefficient storage unit and an output of the second image data selection unit. An image information conversion device characterized by the above-mentioned. 2. The method according to claim 1, wherein the prediction coefficient is determined by a predetermined process based on an image signal having the same signal format as the input image signal and an image signal having the same signal format as the prediction image signal. An image information conversion device characterized by being calculated. 3. The image processing device according to claim 1, wherein the arithmetic processing by the pattern detecting means is performed by calculating a sum of two pixels having phases opposite to each other in the image data selected by the first image data selecting means. 1
An image information conversion apparatus for calculating a value of / 2. 4. The image processing apparatus according to claim 1, wherein the arithmetic processing by the pattern detecting means includes: selecting a pixel having the same phase as the target pixel from among the image data selected by the first image data selecting means; An image information conversion apparatus, which calculates a half value of a difference between a pixel and a pixel having an opposite phase. 5. The image information conversion device according to claim 1, wherein the pattern detection means performs an encoding process adapted to a local dynamic range on a result of the arithmetic process. 6. The image processing apparatus according to claim 1, wherein the second image data selection unit includes: the target pixel; a predetermined number of pixels located around the target pixel in a field including the target pixel; And extracting a predetermined number of pixels including a pixel at a position corresponding to the target pixel in a field other than the field including. 7. The image processing apparatus according to claim 1, wherein the prediction operation processing means includes an offset component generated between the input image signal and the prediction image signal during image information conversion, in addition to an output of the second image data selection means. The image information conversion apparatus according to claim 1, further comprising: 8. The image information conversion device according to claim 1, wherein the input image signal is a composite signal, and the predicted image signal is a component signal. 9. The image information conversion device according to claim 8, wherein the composite signal is an NTSC signal. 10. An image information conversion method for performing image information conversion on an input image signal, comprising: selecting a target pixel and a pixel having a predetermined positional relationship with respect to the target pixel from the input image signal. Performing a predetermined operation related to the phase of the predetermined pixel based on a result of the first image data selection step;
A pattern detection step of detecting a pattern in the vicinity of the pixel of interest based on the result of the arithmetic processing; and a class code generation of generating a class code indicating a class in the vicinity of the pixel of interest based on the result of the pattern detection step A second image data selecting step of selecting a pixel having a predetermined positional relationship with respect to the pixel of interest from the input image signal; and storing a prediction coefficient predetermined for each class, A coefficient storage step of outputting a prediction coefficient corresponding to a result of the code generation step; a prediction calculation processing for performing a predetermined calculation processing based on a result of the coefficient storage step and a result of the second image data selection step And an image information converting method.