JP2004172944A - Image processing method, image processor and program for image processing - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing method capable of improving image quality further for decoded image signals by eliminating mosquito noise generated in compression encoding in decoding and performing an appropriate enhancement processing, an image processor and a program for an image processing. <P>SOLUTION: The decoded image signals for which image signals compressed and encoded by using orthogonal transformation or the like are decoded are separated into low frequency signal components and high frequency signal components. For the high frequency signal components, signals equal to or lower than a prescribed coring level are converted to zero, a prescribed enhancement level is multiplied, the high frequency signal components are composited with the low frequency signal components and the decoded image signals are reconstituted. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an image processing method, an image processing apparatus, and an image processing program for improving image quality of an image obtained by decoding image data that has been compression-encoded such as MPEG. It is another object of the present invention to provide an image processing method, an image processing apparatus, and an image processing program that can remove mosquito noise generated by compression using orthogonal transformation or the like at the time of decoding. .
[0002]
[Prior art]
When transmitting, recording, and reproducing an image signal, an audio signal, and other various signals as digital signals, a compression / expansion technique of an information amount is used. This is because, for example, when digitizing an image signal, an audio signal, or the like, linear quantization (equal quantization) in which each sample value is replaced with one representative value of signal levels obtained by equally dividing the sample value, This is because the amount of information of a signal to be transmitted, recorded, and reproduced becomes extremely large.
[0003]
Conventionally, in the technical field of broadcast communication, in the technical field of recording and reproduction, for example, human vision and hearing are sensitive in parts where signal changes are small, and even if there is some error in parts where signal changes are severe, In addition to using the characteristics of human vision and hearing that are difficult to detect to reduce the amount of information per sample, many information compression technologies have been applied to transmit, record, and reproduce. It is well known that high-efficiency compression technology (high-efficiency encoding technology of information) for various types of information has been put into practical use.
[0004]
By the way, the amount of information per hour in a moving image of a quality similar to a reproduced image displayed using a reproduced signal from a VHS (registered trademark) type VTR currently in practical use is approximately 109 Gbits. In addition, the amount of information per hour in a moving image having an image quality similar to that of a received image of the current standard color television system in Japan is approximately 360 Gbits. A high-efficiency compression method for image information required to transmit, record, and reproduce image information having a larger information amount than these information amounts using current practical transmission lines and recording media. Research on the practical application of has been actively conducted.
[0005]
By the way, in the high-efficiency compression method for image information, which is currently proposed as a high-efficiency compression method for practical image information, (1) the correlation between adjacent pixels in a natural image is high. Compression of the information amount performed by using the correlation (compression of the information amount performed by using the spatial correlation), (2) Performed by using the correlation between screens (inter-frames) arranged on the time axis Combination of three different types of compression means, namely, compression of information amount (compression of information amount by using temporal correlation), and (3) compression of information amount by bias of code appearance probability. In order to achieve highly efficient encoding. Many methods have been proposed as means for compressing the information amount of an image using the above-mentioned correlation (1) in the screen (in the frame). Orthogonal transforms such as -L (Karhunen-Loeve) transform, discrete cosine transform (DCT), discrete Fourier transform, Walsh-Hadamard transform, and the like are often used.
[0006]
For example, a highly efficient encoding method of image information (MPEG1 method, MPEG2 method) proposed as a result of international standardization work by MPEG (Moving Picture Coding Expert Group) established under ISO (International Organization for Standardization). ) Performs high-efficiency encoding of moving image information while performing motion compensation prediction and inter-frame prediction by combining intra-frame encoding and inter-frame encoding. A two-dimensional discrete cosine transform (two-dimensional DCT) is employed as the orthogonal transform. In the orthogonal transform, a predetermined block size (N × M pixels ← horizontal N pixels × vertical M line block size) is calculated for each screen image signal to be subjected to high-efficiency encoding. An image divided for each “unit block” (in the above-described MPEG1 system and MPEG2 system, a block having a block size of 8 × 8 pixels ← 8 horizontal pixels × 8 vertical lines is referred to as a “unit block”). Performed on signals.
[0007]
(N × M) orthogonal transform coefficients (8 × 8 = 64 DCT transform coefficients in the MPEG1 and MPEG2 systems) obtained by orthogonally transforming each block of the unit are at least the unit "Block quantization width value" set for each area of a predetermined size including one of the blocks (the area referred to by the term "macroblock" in the MPEG1 and MPEG2 systems). Is quantized. The “macro block” is an MPEG1 system or an MPEG2 system, which is an area having a block size of 16 × 16 pixels (16 horizontal pixels × 16 vertical lines) for the luminance signal Y and two color difference signals Cr, Cb. This is an area composed of a block size area of 8 × 8 pixels (8 horizontal pixels × 8 vertical lines) for each.
[0008]
For example, in the MPEG1 system and the MPEG2 system, the “block quantization width value” is indicated as [macroblock quantization characteristic value (or quantization scale of macroblock) QS × quantization matrix].
[0009]
The orthogonal transform coefficient (for example, DCT coefficient) quantized by the block quantization width value is separated into its DC component (DC component) and AC component (AC component). The DC component of the orthogonal transform coefficient (for example, DCT coefficient) is differentially coded, and the AC component of the orthogonal transform coefficient (for example, DCT coefficient) is subjected to zigzag scanning and then entropy-encoded (information based on the bias of the code appearance probability). Amount compression: variable-length encoding such as the Huffman method). The image data thus transformed and coded is output as a bit stream (bit string).
[0010]
Next, the decoding operation on the image data transformed and coded as described above is performed in the reverse operation of the above-described coding operation to obtain an output image. When quantization is performed, quantization noise is generated in the output image due to the unavoidable quantization error. Then, when the complexity of the image to be coded is large relative to the transmission rate, the quantization noise greatly degrades the quality of the image. FIG. 9 shows a conventional general MPEG encoder, and FIG. 10 shows a decoder.
[0011]
The conventional MPEG encoder shown in FIG. 9 will be briefly described. The difference between the input image and the image decoded by the motion compensation predictor 1 and the input image is calculated by a differentiator 2 to reduce a time redundant part.
[0012]
The prediction direction has three modes from the past, the future, and both. These can be switched and used for each MB (macroblock) of 16 pixels × 16 pixels. The prediction direction is determined by the picture type given to the input image. A P-picture has two modes: a mode for encoding based on prediction from the past and a mode for independently encoding the MB without performing prediction. A B picture has four modes for prediction from the future, prediction from the past, prediction from both, and independent encoding. An I-picture independently encodes all MBs.
[0013]
In motion compensation, a motion vector is detected with half-pel accuracy by performing pattern matching for each MB, and is predicted after shifting by the amount of motion. The motion vector has a horizontal direction and a vertical direction, and is transmitted as additional information of the MB together with an MC (Motion Compensation) mode indicating where to predict the motion vector.
[0014]
A GOP (Group Of Picture) is referred to as a GOP (Group Of Picture) from an I picture to a picture before the next I picture. In general, about 15 pictures are used as one GOP.
[0015]
The orthogonal transformation is performed on the difference image in the DCT unit 3. DCT (Discrete Cosine Transform) is an orthogonal transform that performs discrete transform of an integral transform using a cosine function as an integral kernel into a finite space. In the MPEG, an MB is divided into four parts, and two-dimensional DCT is performed on an 8 × 8 DCT block. In general, since a video signal has many low-frequency components and few high-frequency components, when DCT is performed, coefficients concentrate on low frequencies.
[0016]
The DCT-processed image data (DCT coefficients) is quantized by a quantizer 4. In the quantization, a value obtained by multiplying an 8 × 8 two-dimensional frequency called a quantization matrix by a visual characteristic by a value called a quantization scale for multiplying the whole by a scalar is used as a quantization value, and the DCT coefficient is quantized. Calculate by value. When inverse quantization is performed by an MPEG decoder (decoder), a value close to the original DCT coefficient is obtained by multiplying by the quantization value.
[0017]
The quantized data is variable-length coded by the VLC unit 5. The direct current (DC) component of the quantized value uses DPCM (differential pulse code modulation) which is one of predictive coding. Huffman coding is performed on the alternating current (AC) component by performing a zigzag scan from a low band to a high band, using a run length of zero and an effective coefficient value as one event, and assigning a code having a short code length from a code having a high appearance probability. Is performed.
[0018]
The variable-length coded data is stored in a temporary buffer 6 and output as coded data at a predetermined transfer rate. The generated code amount of the output data for each macro block is transmitted to the code amount controller 21, and the error code amount between the generated code amount and the target code amount is fed back to the quantizer 4 to quantize the data. The code amount is controlled by adjusting.
[0019]
The quantized image data is inversely quantized by an inverse quantizer 7, inverse DCT by an inverse DCT unit 8, temporarily stored in an image memory 10 via an adder 9, and then stored in a motion compensation predictor 1. Is used as a reference decoded image for calculating a difference image.
[0020]
FIG. 10 shows an MPEG decoder (decoder) for decoding the coded data that has been MPEG-coded in this way.
[0021]
Incoming encoded data (stream) is buffered in a buffer 11, and data from the buffer 11 is input to a VLD unit 12. The VLD unit 12 performs variable length decoding to obtain a direct current (DC) component and an alternating current (AC) component. The alternating current (AC) component data is arranged in an 8 × 8 matrix in the order of zigzag scan from the low band to the high band. This data is input to the inverse quantizer 13 and is inversely quantized by the quantization matrix. The inversely quantized data is input to the inverse DCT unit 14, subjected to inverse DCT, and output as image data (decoded data). The decoded data is temporarily stored in the image memory 16 and then used by the motion compensation predictor 17 as a reference decoded image for calculating a difference image.
[0022]
Generally, among the quantization errors that cause the above-described quantization noise, the quantization error of the low-frequency component causes output image distortion in a state where there is no correlation between unit blocks, so-called block distortion, in the image. Also, among the quantization errors that cause quantization noise, the quantization error of the high-frequency component causes ringing-like output image distortion, so-called mosquito noise, around the edge. By the way, as described above, quantization noise generated in an image is particularly conspicuous in a flat portion of the image. That is, in the case of a quantization noise waveform in which a small noise is added at a place where a large video signal level changes from a low band to a high band, the noise difference is small due to a small sensitivity difference in visual characteristics. Is hard to detect. However, when there is a large change in the video signal level only in the low band, and when a small noise is added in the high band, the noise is easily detected. As a matter of course, when a large noise is added, a fatal error is detected as coding degradation regardless of the low frequency band and the high frequency band.
[0023]
As a countermeasure against the above-described quantization noise, as one of measures conventionally adopted, it is assumed that a portion having a low high-frequency component in a decoded image is noise or quantization noise of an image signal, There is a coring technique in which a signal in which a signal level of an image signal is equal to or lower than a predetermined signal level is regarded as a zero signal portion. It is well known that the above-described coring technique has been used as a method for removing minute noise from analog signals for a long time (see, for example, February 141, "Broadcasting Technology" pages 141 to 147). ).
[0024]
Also, Japanese Patent Application Laid-Open No. 8-205158, which is an invention by the same inventor as the present application, discloses the following method. First, the block quantization width information of the additional information detected by the block quantization width value detection unit from the decoded image data obtained by decoding the bit stream based on the transform-coded image data and the additional information, The pixel address obtained from the memory is supplied to a coring level determination unit to generate a coring level control signal. By changing the coring level of the coring circuit based on the block quantization width information and the pixel address, the coring level increases in areas where the block quantization width value is large, so that mosquito noise is reduced while preserving edges. Is easily performed. In addition, a coring operation with a large coring level is performed on pixels near the boundary of the unit block to effectively reduce mosquito noise and block distortion.
[0025]
[Patent Document 1]
JP-A-8-205158
[0026]
[Problems to be solved by the invention]
In image coding with quantization such as MPEG, if the degree of complexity of an image to be coded with high efficiency is large with respect to the transmission rate, image deterioration is always large. By the way, in the weighting matrix in consideration of the two-dimensional visual characteristic, since the block quantization width value corresponding to the high frequency component is basically set to be larger, quantization noise tends to appear on the high frequency side. Therefore, the decoded image signal is separated into a low-frequency signal component and a high-frequency signal component, and the separated high-frequency signal component is subjected to coring means to convert a signal level below a predetermined signal level (coring level) to zero. It is effective to combine the low-frequency signal component and the correlated high-frequency signal component and output the combined signal.
[0027]
However, the deteriorated image described above often has a relatively large quantization width. Therefore, even if only mosquito noise, which is a high-frequency noise, is removed by coring, the image quality cannot be substantially improved. I can't. That is, the image quality is degraded in the direction in which the entire image is blurred due to the quantization, and improvement has been required.
[0028]
On the other hand, in order to improve the image quality, there is an enhancement technique such as increasing the resolution by performing edge enhancement on the decoded image signal itself.However, an enhancement filter or the like dedicated to the enhancement processing is separately required, so that the number of processing increases and the cost increases. There was a problem that led to an increase.
[0029]
The present invention is directed to an image processing method, an image processing apparatus, and an image processing method capable of removing mosquito noise generated by compression encoding at the time of decoding and performing appropriate enhancement processing to further improve the image quality of a decoded image signal. The purpose is to provide a processing program.
[0030]
[Means for Solving the Problems]
Therefore, in order to solve the above problems, the present invention provides the following image processing method, image processing apparatus, and image processing program.
(1) A decoded image signal obtained by performing a decoding signal process on image data encoded by a predetermined encoding method using orthogonal transform is separated into a low-frequency signal component and a high-frequency signal component, The high-frequency signal component is converted to a signal having a predetermined coring level or less to zero, and then multiplied by a predetermined enhancement level, and the low-frequency signal component and the multiplied high-frequency signal component of the enhancement level are combined. An image processing method for obtaining a decoded image signal from which noise has been removed.
(2) Separation for separating a decoded image signal obtained by performing decoding signal processing on image data encoded by a predetermined encoding method using orthogonal transform into a low-frequency signal component and a high-frequency signal component Means,
Coring means for converting a signal having a predetermined coring level or less to zero for the separated high-frequency signal component,
Enhancement means for multiplying the correlated high-frequency signal component by a predetermined enhancement level,
Combining means for combining the low-frequency signal component and the high-frequency signal component output from the enhancing means,
An image processing apparatus comprising:
(3) Separation for separating a decoded image signal obtained by performing decoding signal processing on image data encoded by a predetermined encoding method using orthogonal transformation into a low-frequency signal component and a high-frequency signal component Features and
For the separated high-frequency signal component, a coring function for converting a signal having a predetermined coring level or less to zero,
An enhancement function of multiplying the correlated high-frequency signal component by a predetermined enhancement level,
A combining function of combining the low-frequency signal component and the enhanced high-frequency signal component;
Image processing program for causing a computer to execute.
[0031]
BEST MODE FOR CARRYING OUT THE INVENTION
The specific contents of the image processing apparatus to which the embodiment of the present invention is applied will be described in detail with reference to FIGS. The encoded data supplied to the input terminal 1 shown in FIG. 1 is obtained by compressing the amount of information of an image by orthogonal transformation using intra-frame (intra-frame) correlation (the amount of information performed using spatial correlation). Of information, compression of information amount using correlation between screens (frames) arranged on the time axis (compression of information amount using temporal correlation), and code appearance probability And image data (for example, image data according to the MPEG1 system and the MPEG2 system) that have been subjected to high-efficiency conversion coding by combining three types of different compression means, that is, compression of the information amount due to the deviation. In the following description of the present specification, description is made assuming that image data to be decoded is image data according to the MPEG1 system and the MPEG2 system.
[0032]
By the way, high-efficiency coding of moving image information in the MPEG1 system and the MPEG2 system combines intra-frame coding with two-dimensional discrete cosine transform (two-dimensional DCT) and inter-frame coding to perform motion compensation prediction and inter-frame coding. It is performed in a state where prediction has been made. Then, the image signal of each screen, which is a target of high-efficiency encoding, is divided into “unit blocks” having a block size of 8 × 8 pixels (horizontal 8 pixels × vertical 8 lines). DCT is performed for each unit block. Then, each of the 64 DCT transform coefficients for each block of each unit is quantized by a “block quantization width value”.
[0033]
In the MPEG1 system and the MPEG2 system, the “block quantization width value” refers to a “macroblock quantization characteristic value (or a macroblock quantization scale) set for each area referred to by the term“ macroblock ”. ) QS ”and“ quantization matrix ”. Here, the “macroblock” is an area of a block size of 16 × 16 pixels (16 horizontal pixels × 16 vertical lines) for the luminance signal Y and 8 × 16 pixels for each of the two color difference signals Cr and Cb. This is an area composed of an area having a block size of 8 pixels (8 horizontal pixels × 8 vertical lines).
[0034]
The DCT coefficient quantized using the DCT transform coefficient as a dividend and the “block quantization width value” as a divisor is separated into its DC component (DC component) and AC component (AC component). The DC component of the DCT coefficient is differentially coded, and the AC component of the DCT coefficient is subjected to zigzag scanning and then entropy coding (information compression by biasing the code appearance probability ... a variable length code such as the Huffman method). ). The image data thus transformed and encoded has additional information (for example, a macroblock quantization characteristic value (or a macroblock quantization scale) which is block quantization width information) required when decoding the image data. QS, motion vector, prediction mode information, etc.) are added to form a bit stream.
[0035]
In FIG. 1, the variable-length decoding unit (corresponding to the block 12 shown in FIG. 10 and the same hereinafter) in the MPEG decoder 32 (the same configuration as that shown in FIG. 10) to which this bit stream is supplied is entropy. Encoded (variable-length encoded) image data and additional information (for example, a macroblock quantization characteristic value (or a macroblock quantization scale that is block quantization width information) required when decoding the image data. ) QS, motion vector, prediction mode information, etc.).
[0036]
The image data decoded by the variable length decoding unit 12 and the block quantization width information are supplied to the inverse quantization unit 13. Further, the motion vector, the prediction mode information, and the like are supplied to the motion compensation prediction unit 17. The inverse quantization unit 13 to which the image data decoded by the variable length decoding unit 1 and the block quantization width information in the decoded additional information are given, converts the DCT transform coefficient obtained by performing the inverse quantization operation. It is supplied to an inverse orthogonal transform unit (inverse DCT) 14.
[0037]
The inverse orthogonal transform unit (inverse DCT) 14 performs two-dimensional inverse DCT for each unit block, inversely transforms image data in the frequency domain into image data in the time domain, and supplies it to the adding unit 15. I do. The image data in the time domain supplied to the addition unit 15 is added to the image data in a state where the motion compensation is performed by the motion compensation prediction unit 13 according to a coding type indicating a difference between intra-frame coding and inter-frame coding. Alternatively, the image data is stored in the image memory 16 as output image data by adding or not adding. The above is the description of the operation of the MPEG decoder 32 shown in FIG.
[0038]
The output of the MPEG decoder 32 shown in FIG. 1 corresponds to FIG. This image data is input to the two-dimensional LPF 33 shown in FIG. The two-dimensional LPF 33 is, for example, a 3 × 3 two-dimensional filter, and multiplies the corresponding pixel data of the image data by a coefficient of {1, 2, 1, 2, 4, 2, 1, 2, 1}, and a multiplication result is obtained. Is divided by 16. By applying this filter, image data of a low-frequency signal component of the image is obtained. It corresponds to FIG.
[0039]
Next, the image data of the low-frequency signal component is subtracted from the output image data of the MPEG decoder 32 by the arithmetic unit 34, thereby separating the high-frequency signal component of the output image data of the MPEG decoder 32. The separated high frequency component image data has positive and negative values. This corresponds to FIG.
[0040]
The image data of the high-frequency signal component is input to the coring circuit 35 shown in FIG. A parameter called a coring level input by a user to an MMI (man machine interface circuit) 36 is transmitted to the coring circuit 35 via the CPU 37.
[0041]
The coring level is a parameter of ± CL centering on 0 as shown in FIG. Image processing by coring is a process of converting a signal having a value of ± CL or less into 0. Since a signal having a low level of a high-frequency component in the decoded image data is often mosquito noise generated at the time of compression, noise can be removed by the coring processing.
[0042]
However, some images have little mosquito noise depending on the content. Therefore, as in the present embodiment, the coring level may be set from the external MMI 36 so that coring according to the amount of mosquito noise can be performed. As an example of the setting, as shown in FIG. 4, the levels 1, 2, and 3 set by the user are defined to correspond to the coring levels (CL) of 2, 4, and 6, respectively. The larger the value, the more the coring works. This corresponds to FIG.
[0043]
Next, the image data high-frequency signal component subjected to the coring processing by the coring circuit 35 is input to the enhancement circuit 38. The enhancement circuit 38 removes the low-level noise of the image data high-frequency signal by the coring circuit 35. Amplify components. As shown in FIG. 8, for example, assuming that the gain between the maximum value and the minimum value is 1.0, as shown in FIG. 8, the gain is set to EL times. EL is an enhancement level.
[0044]
A parameter called a coring level input by the user to the MMI 36 is transmitted to the enhancement circuit 38 via the CPU 37. The enhancement level is set in such a manner that levels 1, 2, and 3 set by the user as shown in FIG. 5 correspond to the enhancement level (EL) values of 1.125, 1.1875, and 1.25, respectively. . The data thus enhanced corresponds to FIG.
[0045]
The enhanced image data high-frequency signal component is added to the image data of the image data low-frequency signal component, which is output data of the two-dimensional LPF 32, by the arithmetic unit 39, and is output as a reconstructed decoded image.
[0046]
FIG. 2 is an example in which the two-dimensional LPF of this image processing apparatus is replaced with horizontal and vertical one-dimensional filters. The one-dimensional LPF is a filter that multiplies the corresponding pixel data of the image data by, for example, a coefficient of {1, 2, 1} and divides the calculation result by four. When this filter is applied, image data of a low-frequency signal component of the image is obtained. The other configuration is the same as that shown in FIG. The one-dimensional configuration requires less arithmetic processing. However, although the two-dimensional configuration increases the product-sum operation processing amount, it has the advantage of being able to handle oblique bands.
[0047]
Thus, in the above embodiment, a decoded image signal obtained by decoding an image signal compressed and encoded using an orthogonal transform or the like is separated into a low-frequency signal component and a high-frequency signal component. A signal below a predetermined coring level is converted to zero, multiplied by a predetermined enhancement level, and its high-frequency signal component is combined with the low-frequency signal component to reconstruct a decoded image signal. Therefore, in the above embodiment, the mosquito noise can be effectively removed from the decoded image signal, and the image data after the mosquito noise removal is enhanced, so that the image quality of the decoded image signal can be further improved. realizable. Further, in the above embodiment, since the enhancement processing is performed only on the high-frequency signal component separated from the decoded image signal and cored, the processing amount and the processing amount are significantly greater than when the entire decoded image signal is enhanced. Time and cost can be reduced.
[0048]
Furthermore, since the coring level and the enhancement level can be changed on the user interface, even if the noise level of the mosquito varies depending on the content, the strength of the coring and enhancement can be set according to the user's preference.
[0049]
Next, an algorithm flowchart in an embodiment of the image processing program of the present invention will be described with reference to FIG. First, a coring level and an enhancement level are set by the user from the MMI. Next, the MPEG image is decoded. Next, one-dimensional or two-dimensional LPF is performed to obtain a low-frequency component of the decoded image data. Next, the low frequency components of the decoded image data are subtracted from the decoded image data.
[0050]
Next, it is determined for each pixel whether or not the image data of the high frequency component obtained by the subtraction is pixel data equal to or less than ± coring level. As a result of the determination, the pixel of NO is left as it is, and the pixel data of YES converts the level to 0. Next, the pixel data is multiplied by the enhancement level. Next, it is determined whether or not the pixel data to be processed still exists in the screen. If there is still, the process returns to the block for determining whether the pixel data is equal to or lower than the ± coring level. When all the data in the screen has been processed, the processed image data and the low-frequency component image data are added.
[0051]
Next, it is determined whether there is still an MPEG encoded image. If there is, the process returns to the block for MPEG decoding. If there is no more MPEG encoded image, the process ends. Using such a flow, a program having an image processing function of the present invention can be realized.
[0052]
Although the present embodiment has been described as an image processing apparatus, it is of course possible to use the apparatus as a reproducing apparatus for a medium on which encoded data such as MPEG is recorded or a receiving apparatus of a system in which encoded data is transmitted. It is.
[0053]
【The invention's effect】
As is clear from the above description, the present invention separates a decoded image signal obtained by decoding an image signal compressed and encoded by using an orthogonal transform or the like into a low-frequency signal component and a high-frequency signal component. For the high-frequency signal component, a signal lower than a predetermined coring level is converted to zero, multiplied by a predetermined enhancement level, and the high-frequency signal component is combined with the low-frequency signal component to obtain a decoded image signal. Reconfigured. Therefore, according to the present invention, the mosquito noise can be effectively removed from the decoded image signal, and the image data after the mosquito noise removal is enhanced, so that the image quality of the decoded image signal is further improved. it can. In addition, since the present invention is an enhancement process only for a high-frequency signal component separated from a decoded image signal and cored, when the entire decoded image signal is enhanced (for example, the enhancement process is performed by separately providing a high-pass filter or the like). The processing amount, processing time, and cost can be greatly reduced as compared with the case of processing.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration example of an apparatus that realizes an image processing apparatus to which an embodiment of the present invention is applied by using a two-dimensional LPF.
FIG. 2 is a block diagram illustrating a configuration example of an apparatus that realizes an image processing apparatus to which an embodiment of the present invention is applied by using a one-dimensional LPF.
FIG. 3 is a diagram illustrating an image processing operation according to an embodiment;
FIG. 4 is an explanatory diagram showing a relationship between a user setting value and a coring level according to one embodiment.
FIG. 5 is an explanatory diagram showing a relationship between a user setting value and an enhancement level in one embodiment.
FIG. 6 is a flowchart of an image processing algorithm according to one embodiment.
FIG. 7 is an explanatory diagram showing an operation of coring according to one embodiment.
FIG. 8 is an explanatory diagram showing an enhancement operation of one embodiment.
FIG. 9 is a block diagram of a conventional MPEG encoder.
FIG. 10 is a block diagram of a conventional MPEG decoder.

Claims (3)

A decoded image signal obtained by performing decoding signal processing on image data encoded by a predetermined encoding method using an orthogonal transform is separated into a low-frequency signal component and a high-frequency signal component, and the high-frequency signal After converting a signal below a predetermined coring level to zero, the component is multiplied by a predetermined enhancement level, and the low-frequency signal component and the high-frequency signal component multiplied by the enhancement level are combined. An image processing method for obtaining a decoded image signal from which noise has been removed. Separation means for separating a decoded image signal obtained by performing decoded signal processing on image data encoded by a predetermined encoding method using an orthogonal transform into a low-frequency signal component and a high-frequency signal component,
Coring means for converting a signal having a predetermined coring level or less to zero for the separated high-frequency signal component,
Enhancement means for multiplying the correlated high-frequency signal component by a predetermined enhancement level,
Combining means for combining the low-frequency signal component and the high-frequency signal component output from the enhancing means,
An image processing apparatus comprising: A separation function of separating a decoded image signal obtained by performing a decoding signal process on image data encoded by a predetermined encoding method using an orthogonal transform into a low-frequency signal component and a high-frequency signal component,
For the separated high-frequency signal component, a coring function for converting a signal having a predetermined coring level or less to zero,
An enhancement function of multiplying the correlated high-frequency signal component by a predetermined enhancement level,
A combining function of combining the low-frequency signal component and the enhanced high-frequency signal component;
Image processing program for causing a computer to execute.

Images