JP4696412B2 - Image decoding device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image decoding apparatus suitable for use when decompressing and decoding image information that has been compression-encoded using, for example, an MPEG (Moving Picture Coding Experts Group) system. Specifically, for example, in an image information compression / decompression system such as the MPEG system, the reduction characteristic against distortion generated according to the bit rate is made variable so that good image information can always be decoded. is there.
[0002]
[Prior art]
For example, in an MPEG image information compression / decompression system, if the bit rate at the time of compression is insufficient, high-frequency components of image information are reduced in the compression process. In such a case, in the decompressed and decoded image information, block distortion occurs at the boundary of the compressed block. As countermeasures, for example, a method of increasing the bit rate to reduce the amount of reduction of high-frequency components, or a method of performing block distortion reduction processing for reducing block distortion on decompressed and decoded image information can be considered.
[0003]
However, in these countermeasures, the method for increasing the bit rate is difficult to implement because it involves a system change or the like. For this reason, conventionally, a method of performing block distortion reduction processing for reducing block distortion on the decompressed and decoded image information is generally employed. However, in the conventional block distortion reduction process, the determination standard of the block distortion reduction process used is always constant, and the effect of improving the image quality is low for low bit rate image information such as an MPEG search image. It was a thing.
[0004]
[Problems to be solved by the invention]
This application has been made in view of the above points, and the problem to be solved is that, in the conventional apparatus, the determination criterion of the block distortion reduction processing used is always constant, for example, This means that good image quality improvement effects could not be obtained for low bit rate image information.
[0005]
[Means for Solving the Problems]
For this reason, in the present invention, the threshold value for determining the correction amount is increased when the bit rate is low, so that the range of the normal correction amount is widened and small when the bit rate is high. According to the present invention, it is possible to obtain a good image quality improvement effect for low bit rate image information, and there is no excessive correction for high bit rate image information. An optimal image quality improvement effect can always be obtained for bit rate image information.
[0006]
DETAILED DESCRIPTION OF THE INVENTION
That is, in the present invention, an image decoding apparatus that decompresses and decodes input image information encoded and compressed for each arbitrary block,
With respect to the decoded image information, as a process of extracting the feature of the boundary of the compressed block, a process of obtaining the unevenness information of the pixel of the boundary of the compressed block and a process of obtaining the absolute value of the step of the boundary of the compressed block Block boundary feature extraction means to perform;
Whether at least the absolute value of the step is equal to or less than a first threshold as a process for determining whether or not the decoded image information should be corrected as block distortion based on the extraction result of the block boundary feature extraction means If it is not less than or equal to the first threshold, it is determined that the block distortion is not, and if the absolute value of the step is less than or equal to the first threshold, the absolute value of the step is A block distortion determination means for determining whether or not the value of the information is greater than or equal to the information value of the unevenness and determining that it is not a block distortion;
As a process for calculating the correction amount of block distortion, when it is determined that the block distortion is determined by the block distortion determination means, a process for calculating a basic correction amount based on the extraction result of the block boundary feature extraction means , it is determined whether the absolute value of the previous SL step is equal to or less than the second threshold value is less than or equal the second threshold value is determined as it is as the correction amount a correction amount of the basic, if not below the second threshold Determines whether or not the absolute value of the step is less than or equal to a third threshold value. If the absolute value of the step is less than or equal to the third threshold value, half of the basic correction amount is determined as the correction amount, A block distortion correction strength calculating means for performing a process of determining a quarter of the basic correction amount as a correction amount when not below the threshold ;
Said block distortion determination unit, wherein the block distortion correction intensity calculation unit, before Symbol first threshold, the second threshold and the third threshold value as 0 ≦ a second threshold value ≦ third threshold ≦ first threshold value The first threshold value can be arbitrarily set, and for the low bit rate image information, the second threshold value and the third threshold value are set higher than the high bit rate image information . Control means for increasing the threshold ;
Block distortion reduction means for reducing block distortion generated at the boundary of the compressed block using the block distortion correction amount calculated by the block distortion correction intensity calculation means for the decoded image information. Is.
[0007]
Hereinafter, the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an embodiment of a block distortion reducing unit which is a main part of an image decoding apparatus to which the present invention is applied.
[0008]
In FIG. 1, for example, decompressed image information (8 bits) is input to the input terminal 1. The image information from the input terminal 1 is supplied to a block boundary feature extraction unit 2 that extracts, for example, a boundary feature between adjacent 8 × 8 pixel compressed blocks. The block boundary feature extraction unit 2 calculates block boundary activities, boundary steps, and the like, and supplies them to a block distortion determination unit 3 for determining whether or not the calculation result is a subsequent block distortion.
[0009]
Further, the block distortion determination unit 3 uses the image information based on the activity and boundary step supplied from the block boundary feature extraction unit 2 and an arbitrarily set threshold 1 (8 bits) supplied from the external controller unit 100. Is to be corrected as block distortion. Then, the determination result in the block distortion determination unit 3 is supplied to the block distortion correction intensity calculation unit 4 for calculating the block distortion correction level in the subsequent stage.
[0010]
Further, the block distortion correction intensity calculation unit 4 uses a pixel around the block boundary, a step at the block boundary, and a threshold value 2 and a threshold value 3 (8 bits each) supplied from the controller unit 100, and the like. A distortion correction amount is calculated. Then, the block distortion correction amount calculated by the block distortion correction intensity calculator 4 is supplied to the block distortion corrector 5 for actually correcting the block distortion in the subsequent stage.
[0011]
Further, the block distortion correction unit 5 adds or subtracts the above-described block distortion correction amount with respect to pixels around the block boundary of the image information supplied from the input terminal 1 so that the block distortion is reduced. Is. Further, the image information whose block distortion has been reduced by the block distortion correction unit 5 is selected as image information from the input terminal 1 by the switch 6 and is taken out to the output terminal 7. In this way, image information with reduced block distortion is taken out to the output terminal 7.
[0012]
In this apparatus, the specific block feature extraction operation in the block boundary feature extraction unit 2 is as follows. That is, FIG. 2 shows, for example, block boundaries and correction ranges for adjacent compressed blocks of 8 × 8 pixels. With respect to these pixels, information on the unevenness of the pixels (activity: hereinafter abbreviated as “act”) for five pixels before and after the block boundary is examined by the following equation.
act = (| b−a | + | c−b | + | d−c | + | e−d | + | g−f | + | h−g | + | i−h | + | j−i | ) / 8
[0013]
Further, an absolute value (hereinafter abbreviated as atmp0) of the step of the compressed block boundary of 8 × 8 pixels is obtained by the following equation.
atmp0 = | fe |
[0014]
Further, these pieces of information are supplied to the subsequent block distortion determination unit 3. The specific block distortion determination operation in the block distortion determination unit 3 is performed as follows. That is, FIG. 3 shows a block distortion determination algorithm. In FIG. 3, when the operation is started, it is determined in step [1] whether or not the absolute value [atmp0] is equal to or less than an arbitrarily set threshold value 1. If [atmp0 ≦ threshold 1] is not satisfied (N), it is determined in step [2] that there is no block distortion, and the process is terminated.
[0015]
When [atmp0 ≦ threshold 1] is satisfied in step [1] (Y), it is determined in step [3] whether or not the absolute value [atmp0] is greater than or equal to the activity [act]. If [atmp0 ≧ act] is not satisfied (N), it is determined in step [2] that there is no block distortion, and the process is terminated.
[0016]
Further, when [atmp0 ≧ act] is satisfied in step [3] (Y), it is determined in step [4] whether or not the absolute value [atmp0] is larger than the value [| ef−]. If it is not [atmp0> | e−f |] (N), it is determined in step [2] that there is no block distortion, and the process ends.
[0017]
When [atmp0> | e−f |] is satisfied in step [4] (Y), it is determined in step [5] whether or not the absolute value [atmp0] is larger than the value [| g−f |]. . If it is not [atmp0> | e−f |] (N), it is determined in step [2] that there is no block distortion, and the process ends. When all of the above steps [1] and steps [3] to [5] are (Y), it is determined in step [6] that block distortion has occurred, and the process ends.
[0018]
That is, in the above-described determination, step [1] uses the feature that when the step at the boundary is large, it is not the step formed by block compression but is often the step of the original image. Here, the determination value can be arbitrarily changed by setting the threshold 1.
[0019]
Step [3] utilizes the feature that the boundary step is larger than the unevenness of the peripheral pixels in the case of block distortion. Further, Steps [4] and [5] utilize the feature that the boundary step is larger than the steps at both ends of the boundary in the case of block distortion.
[0020]
For the block distortion determined in this way, the block distortion correction intensity calculation unit 4 calculates a basic correction amount of block distortion (hereinafter abbreviated as sigma0) by the following equation.
sigma0 = | atmp0− {| (ed −) + (g−f) | / 2} | / 2
[0021]
Then, for the basic correction amount [sigma0] of the block distortion calculated in this way, the correction amount is corrected as shown in FIG. In this operation, the relationship between threshold 1, threshold 2, and threshold 3 is 0 ≦ threshold 2 ≦ threshold 3 ≦ threshold 1.
[0022]
That is, in FIG. 4, when the operation is started, it is determined in step [11] whether block distortion is present. Here, the determination of whether or not the block distortion is caused is based on the above-described step [2] or [6] in FIG. If it is not block distortion (N), the correction amount is set to 0 in step [12] and the operation is terminated.
[0023]
When block distortion is detected in step [11] (Y), it is determined in step [13] whether or not the absolute value [atmp0] is equal to or less than a threshold value 2 that is arbitrarily set. When [atmp0 ≦ threshold 2] is satisfied (Y), the correction amount = [sigma0] is set in step [14] and the operation is terminated.
[0024]
Further, when [atmp0 ≦ threshold 2] is not satisfied in step [13] (N), it is determined in step [15] whether or not the absolute value [atmp0] is equal to or less than a threshold value 3 that is arbitrarily set. Here, when [atmp0 ≦ threshold 3] is satisfied (Y), the correction amount = [sigma0 / 2] is set in step [16] and the operation is terminated. If [atmp0 ≦ threshold 3] is not satisfied in step [15] (N), the correction amount is set to [sigma0 / 4] in step [17] and the operation is terminated.
[0025]
In this way, the correction amount is determined. In this case, since the threshold value 2 and the threshold value 3 can be set from outside, optimum parameters corresponding to various bit rates can be obtained.
[0026]
That is, when the block boundary is large even when it is determined as block distortion at a high bit rate, it is often a step of the original image rather than the step generated by block compression. Therefore, when the bit rate is high, erroneous correction can be reduced by setting the threshold values 2 and 3 small by external control.
[0027]
Further, the block boundary step at the time of a low bit rate is almost always a step appearing by block compression. Therefore, when the bit rate is low, the threshold value 2 and the threshold value 3 are increased by external control, so that the range of the normal correction amount can be widened so that the image quality can be improved.
[0028]
Further, the correction amount determined in this way is supplied to the block distortion correction unit 5. The block distortion correction unit 5 corrects the block distortion of the input image information using the supplied correction amount. That is, the supplied correction amount itself is used for pixels at the block boundary, 1/2 of the supplied correction amount is used for pixels one pixel apart, and supplied to pixels two pixels apart. Correction is performed so that each of the boundary steps becomes small by using 1/4 of the correction amount to be used and 1/8 of the supplied correction amount for the pixels separated by 3 pixels.
[0029]
In this way, for example, image information having a step at the block boundary as indicated by a circle in FIG. 5 is not corrected for the pixels a and j, and 1/8 of the basic correction amount for the pixels b and i. , For the pixels c and h, there is a boundary step as ¼ of the basic correction amount, for the pixels d and g, ½ of the basic correction amount, and for the pixels e and f, as the basic correction amount. By adding and subtracting the correction amount so as to decrease, correction for reducing block distortion is performed, for example, as indicated by Δ in FIG.
[0030]
Then, as described above, the threshold value 1, threshold value 2, and threshold value 3 that determine these correction amounts are determined by the external controller unit 100 to be low bit rate image information such as an MPEG search image, for example. By controlling adaptively,
The effect of reducing block distortion can be enhanced for low bit rate image information such as search images.
-For high bit rate image information such as a normal image, failure due to overcorrection can be reduced.
Such effects can be obtained.
[0031]
Therefore, in this embodiment, by changing the characteristics of the block distortion reducing means in accordance with the bit rate of the input image information, it is possible to obtain a good image quality improvement effect for the low bit rate image information. At the same time, there is no excessive correction for image information with a high bit rate, and an optimum image quality improvement effect can always be obtained for image information with various bit rates.
[0032]
As a result, in the conventional apparatus, the determination criterion for the block distortion reduction processing used is always constant, and for example, it has not been possible to obtain a good image quality improvement effect for low bit rate image information. According to the present invention, these problems can be easily solved.
[0033]
Thus, according to the above-described image decoding apparatus, an image decoding apparatus that decompresses and decodes input image information encoded and compressed for each arbitrary block, and blocks image information after decoding. Block distortion reducing means for reducing block distortion generated at the boundary of the image, and by changing the characteristics of the block distortion reducing means according to the bit rate of the input image information, it is favorable for low bit rate image information. The image quality improvement effect can be obtained, and there is no excessive correction for high bit rate image information, and the image quality improvement effect is always optimal for various bit rate image information. Can be obtained.
[0034]
Note that the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit of the present invention.
[0035]
【The invention's effect】
Therefore, according to the first aspect of the present invention, the second threshold value and the third threshold value that determine the correction amount are increased when the bit rate is low, thereby widening the range of the normal correction amount, and when the bit rate is high. By giving it small, it is possible to obtain a good image quality improvement effect for low bit rate image information, and excessive correction is performed for high bit rate image information. In other words, it is possible to always obtain an optimal image quality improvement effect for image information of various bit rates.
[0040]
As a result, in the conventional apparatus, the determination criterion for the block distortion reduction processing used is always constant, and for example, it has not been possible to obtain a good image quality improvement effect for low bit rate image information. According to the present invention, these problems can be easily solved.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of an embodiment of a block distortion reducing unit that is a main part of an image decoding apparatus to which the present invention is applied.
FIG. 2 is a diagram for explaining that;
FIG. 3 is a flowchart for explaining the operation.
FIG. 4 is a flowchart for explaining the operation.
FIG. 5 is a diagram for explaining the effect.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Input terminal, 2 ... Block boundary feature extraction part, 3 ... Block distortion determination part, 4 ... Block distortion correction intensity calculation part, 5 ... Block distortion correction part, 6 ... Switch, 7 ... Output terminal, 100 ... External controller Part

Claims (1)

An image decoding apparatus for decompressing and decoding input image information encoded and compressed for each arbitrary block,
With respect to the decoded image information, as a process of extracting the feature of the boundary of the compressed block, a process of obtaining the unevenness information of the pixel of the boundary of the compressed block and a process of obtaining the absolute value of the step of the boundary of the compressed block Block boundary feature extraction means to perform;
Whether at least the absolute value of the step is equal to or less than a first threshold as a process for determining whether or not the decoded image information should be corrected as block distortion based on the extraction result of the block boundary feature extraction means If it is not less than or equal to the first threshold, it is determined that the block distortion is not, and if the absolute value of the step is less than or equal to the first threshold, the absolute value of the step is A block distortion determination means for determining whether or not the value of the information is greater than or equal to the information value of the unevenness and determining that it is not a block distortion;
As a process for calculating the correction amount of block distortion, when it is determined that the block distortion is determined by the block distortion determination means, a process for calculating a basic correction amount based on the extraction result of the block boundary feature extraction means , it is determined whether the absolute value of the previous SL step is equal to or less than the second threshold value is less than or equal the second threshold value is determined as it is as the correction amount a correction amount of the basic, if not below the second threshold Determines whether or not the absolute value of the step is less than or equal to a third threshold value. If the absolute value of the step is less than or equal to the third threshold value, half of the basic correction amount is determined as the correction amount, A block distortion correction strength calculating means for performing a process of determining a quarter of the basic correction amount as a correction amount when not below the threshold ;
Said block distortion determination unit, wherein the block distortion correction intensity calculation unit, before Symbol first threshold, the second threshold and the third threshold value as 0 ≦ a second threshold value ≦ third threshold ≦ first threshold value The first threshold value can be arbitrarily set, and for the low bit rate image information, the second threshold value and the third threshold value are set higher than the high bit rate image information . Control means for increasing the threshold ;
Block distortion reduction means for reducing block distortion generated at the boundary of the compressed block using the block distortion correction amount calculated by the block distortion correction intensity calculation means for the decoded image information. Image decoding device.

Images