JP2000050262A - Device and method for image decoding, and device and method for image coding - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize an image with high image quality with less blurring in the case that magnification is conducted for decoding the image. SOLUTION: This decoder has a reference image decoding section 1 which acts as an initial image generating section that receives coded information to generate an image which is a source of a decoding image, a polygonal shape image conversion section 2a that converts a polygonal image into other polygonal image in each image generated at each stage, and an adder section 3a acting like a compositing section that composites the image before the conversion and the image after the conversion. The polygonal shape image conversion section 2a and the adder section 3a adopt a multi-stage configuration. The polygon shape image conversion section 2a and the adder section 3a being the compositing section configure a 1st stage, a polygon shape image conversion section 2b and an adder section 3b configure a 2nd stage respectively, and each stage is similarly added to configure a multi-stage.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image decoding apparatus and method, and an image encoding apparatus and method, and more particularly, to an image decoding apparatus and method for realizing decoding that can obtain a high-quality restored image even if the resolution is changed. TECHNICAL FIELD The present invention relates to a method, and an image encoding device and method.

[0002]

2. Description of the Related Art As a conventional typical image compression method, I
The so-called JPEG standardized by SO (Joint Ph
The otographic Coding Experts Group) method is known. This JPEG method uses DCT (discrete cosine transform:
It is known that when a relatively high bit is assigned using Discrete Cosine Transform, good encoded / decoded images are provided. However, if the number of coded bits is reduced to some extent, block distortion peculiar to DCT becomes remarkable, and deterioration is subjectively noticeable.

[0003] Apart from this, recently, researches on a method of dividing an image into a plurality of bands by a filter called a filter bank, which is a combination of a high-pass filter and a low-pass filter, and performing encoding for each band, have been actively conducted. It has become. Among them, wavelet coding uses D
Since there is no disadvantage that block distortion becomes remarkable due to high compression, which is a problem in CT, it is regarded as a promising new technology to replace DCT.

FIG. 13 is a block diagram showing a typical wavelet encoding device. The wavelet encoding device shown in FIG. 13 includes a wavelet transform unit 13, a transform coefficient scanning unit 14, a quantization unit 15, and an entropy encoding unit 16.

[0005] In FIG. 13, a wavelet transform unit 13 that has received an input image 117 performs a wavelet transform on the input image to a predetermined number of levels, and outputs a transform coefficient 124 for each level. This wavelet transform will be described later. The wavelet transform coefficient 124 from the wavelet transform unit 13 is input to the transform coefficient scanning unit 14. Transform coefficient scanning unit 14
Then, the rearrangement of the wavelet transform coefficients 124 is performed, and the rearranged wavelet transform coefficients 125 are sent to the quantization unit 15. The quantization unit 15 quantizes the wavelet transform coefficient 125 and sends a quantization coefficient 126 to the entropy coding unit 16. Entropy encoder 16
Generates an encoded bit stream 127 by performing information source encoding such as Huffman encoding or arithmetic encoding based on the quantization coefficient 126, and outputs the encoded bit stream 127.

FIG. 14 is a diagram showing an example of a configuration for performing band division using wavelet transform for each level as the wavelet transform unit 13 in FIG. Wavelet transform section shown in FIG. 14, the circuit unit 20 to the multi-stage (in FIG. 14 20 ₁ serving as a basic unit, 20 _2, 20 ₃
(Three stages).

In FIG. 14, a circuit unit 20 (in FIG. 14, 20 ₁ ,
20 ₂ , 20 ₃ ) are a low-pass filter (low-pass filter: H ₀ (z)) 17, a high-pass filter (high-pass filter: H ₁ (z)) 18, and each of the filters 17, 18. Each of the output signals is configured to have downsamplers 19 _a and 19 _b for reducing the resolution of each output signal to half. As shown in FIG. 14, the input image is first input to the circuit portion 20 ₁ of level 1 is divided into components of the respective low and high band by the low-pass filter 17 and high pass filter 18, each further down-sampler in 19 _a, 19 _b is the resolution in half. Then, at level 2, the components in the low frequency side obtained above, by further circuit 20 _2, downsampling is divided into low-frequency and the high-frequency component. Thereafter, by repeating the same operation up to a predetermined level, band division in which low-frequency components are hierarchically divided can be performed. FIG. 14 shows an example in which the level is up to 3, and 4
Two signals 128, 129, 130, 131 are output.

On the other hand, FIG. 15 is a block diagram showing a configuration example of a wavelet inverse transform unit for performing a wavelet inverse transform which is the inverse of the above wavelet transform. The wavelet inverse transform unit shown in FIG. The circuit unit 25 serving as a basic unit is provided in multiple stages (25 ₁ , 2 in FIG.
5 to _2, 25 ₃ of the third stage) is obtained by construction. The circuit unit 25, which is a basic unit of the inverse wavelet transform, includes a double up-sampler 21 _a , 21 _b for up-sampling the input signal twice, a low-pass filter 22 represented by a transfer function F ₀ (z), It comprises a high-pass filter 23 represented by a transfer function F ₁ (z) and an adder 24.

In the inverse wavelet transform section shown in FIG. 15, for example, the four-waveform obtained by the circuit shown in FIG.
Signals 128, 129, 130 and 131 are input. First, the signal 12 is output by the level 1 circuit section 25 _1.
8 and 129 are upsamplers 21 _a and 21 respectively
_{The signal} is up-sampled twice by _b , sent to the low-pass filter 22 and the high-pass filter 23, and converted into a low-frequency component and a high-frequency component, respectively. Then, the circuit portion 25 ₂ of the level 2, performs addition synthesis process similar upsampled and low-frequency component and a high-frequency component with respect to the signal and the input signal 130 from the circuit unit 25 _1. Thereafter, the same operation is repeated to a predetermined level to obtain a final output signal. FIG. 15 shows an example in which the level is up to 3, and a final output image signal 132 is obtained. When the wavelet transform unit in FIG. 14 and the wavelet inverse transform unit in FIG. 15 are configured using a completely reconfigurable low-pass filter and high-pass filter (filter bank), the output signal 132 is It is equal to 117 of the image signal.

FIG. 16 shows a two-dimensional image (the horizontal size is X,
FIG. 17 shows a division diagram in a case where the vertical size is Y) divided into wavelets in, for example, two stages, and FIG. 17 shows a specific example of actual division using an image.

First, level 1 band division is performed as shown in FIG.
, Four band images LL, LH, HL, and HH are generated. Here, for example, HL means a band component in which the vertical component is high (H) and the horizontal component is low (L). Subsequently, LL is divided by level 2 to LL.
Four band images LL, LLHL, LLLH, and LLHH are generated. In this case, it can be seen from FIG. 17 that most of the information of the image is included in the LLLL band.

When restoring an image using the inverse wavelet transform, the band image may be restored by applying a low-pass filter and a high-pass filter to the signal of each band by the configuration shown in FIG. The above is the description of the conventional wavelet transform and the inverse wavelet transform.

[0013]

When an image having a resolution exceeding the resolution of the input image 117 is restored on the decoding side by using the above-described conventional wavelet transform technique, for example, the resolution of the input image 117 (for example, Consider a case where an attempt is made to obtain a resolution twice the horizontal size of X and the vertical size of Y).

At this time, on the decoding side, (horizontal X, vertical Y)
Is regarded as a low-frequency component (LL), and the rest of the 3
If one band (LH, HL, HH) component is set to 0 and inverse wavelet transform is applied using, for example, the configuration shown in FIG. 15, an enlarged image with (horizontal 2X, vertical 2Y) resolution can be restored.

However, in such a method, since 0 is inserted in the coefficients other than the low-frequency component, the enlarged image is blurred in an image portion having a high-frequency component such as an edge portion, and has poor reproducibility. Tend to be.

The present invention provides an image decoding apparatus and method, and an image encoding apparatus and method in which the blur is small and the components of the original image are reproduced even when the decoding involves enlargement at the time of decoding as described above. The purpose is to provide.

[0017]

According to the present invention, there is provided an image decoding apparatus and method according to the present invention, in which encoding information is input to generate an image which is a basis of a decoded image, and a polygon in an image generated at each stage is generated. The above-mentioned problem is solved by performing a conversion from an image to the other polygon image, synthesizing the image before the conversion and the image after the conversion, and configuring the polygon image conversion and the synthesis in multiple stages. .

In this image decoding, reference image decoding decodes an encoded bit stream to decode a reference image,
The polygon image conversion performs a conversion from a polygon image of a conversion source to a polygon image of a conversion destination based on input encoding information of a polygon image. In addition, the converted image and the image before the conversion are added.

Next, an image coding apparatus and method according to the present invention downsamples an input image and outputs an address between two polygon images having different components in the downsampled image as coding information. Then, the down-sampled image is up-sampled, the up-sampled image is decoded on a polygon image basis based on the encoding information output by the polygon image encoding, and the input image is It is characterized by encoding and outputting encoded information.

In this image coding, the resolution is converted into an image in which the resolution of the original image is reduced by downsampling, and the image is converted into an image in which the resolution of the original image is increased by upsampling. In polygon image coding, a polygon image serving as a conversion source is searched for all polygon images constituting an image, and the position (address) of the polygon image is searched.
Is transmitted as encoded information. Polygon image decoding performs local decoding, performs polygon image conversion based on the obtained encoded information, and provides a decoded image. Further, the original image is encoded by a predetermined encoding method, and an encoded bit stream is transmitted.

[0021]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Some preferred embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an image decoding apparatus according to a first embodiment of the present invention. The image decoding apparatus shown in FIG. 1 includes a reference image decoding unit 1 as an initial image generation unit that receives encoded information and generates an image serving as a source of a decoded image. A polygon image conversion unit 2 for converting a certain polygon image into another polygon image, and an addition unit 3 as a synthesis unit for synthesizing the image before conversion and the image after conversion, The image converter 2 and the adder 3 have a multi-stage configuration. That is, the first stage is constituted by the polygon image conversion unit 2a and the addition unit 3a which is a synthesis unit, and the second stage is constituted by the polygon image conversion unit 2b and the addition unit 3b. Add
It has multiple stages.

Next, the operation will be described. In FIG. 1, a reference image decoding unit 1 that has received an encoded bit stream 100 performs a predetermined decoding process to restore a restored image 10.
1 is generated and output. Subsequently, the first polygon image conversion unit 2 inputs the polygon image encoding information 108 sent from the encoder, and converts the polygon image from the source of the polygon image in the screen of the decoded image 101 to the destination of the polygon image. The information to be converted is decrypted, and this is performed on all the polygon images constituting the restored image 101.

The polygon image converter 2 generated as a result
The converted image 102a is added to the decoded image 101 in the adding unit 3a, and the first added image 10
3 is obtained as a decoded image 4 (I ₁ ). Here, the above-described addition processing includes averaging an image before conversion by the polygon image conversion unit and an image after conversion, or multiplying by a predetermined ratio, and specifically, adding. For example, an average of the decoded image 101 and the converted image 102 may be used, or a multiplication by a predetermined ratio may be added. The above is the first-stage decoding process. By repeating this process up to a predetermined number of stages, a final decoded output is obtained. In FIG. 1, only two stages are shown for simplification of the description. When the decoding process ends at the second stage as shown in FIG. 1, the decoded image 5 which is the added image 106 from the second stage adder 3b.
(I ₂ ) becomes the final decoded image 106, which is extracted as a decoded output of the apparatus in FIG.

Next, the polygon image conversion unit 2 (2 in FIG. 1)
The operation in (a, 2b) will be described. The polygon image conversion unit 2 converts a certain polygon image into another polygon image for each polygon image constituting the entire screen. Position information of the two polygon images on the screen and information on the sizes of the polygon images are all described in the polygon image encoding information 108. As is evident from FIG. 1, the polygon image encoding information 108 is exactly the same at any stage in the polygon image conversion unit 2 for converting one polygon image into another polygon image as described above. Processing is performed.

FIG. 2 shows a specific process of the polygon image conversion.
This will be described with reference to FIG. As the contents of the polygon image encoding information 108, only the position information (specifically, the address) of the polygon image of the conversion destination and the polygon image of the conversion source on the screen may be sufficient.
This is effective when a certain pixel value in the source polygon image is copied as it is as the pixel value of the destination polygon image.

In FIG. 2, A _k indicates a polygon image of a conversion destination, and B _k indicates a polygon image of a conversion source.
Numbers from “1” to “10” in a circle in the figure each represent a pixel, and pixels in the polygon image B _k of the conversion source are converted to pixels in the polygon image A _k of the conversion destination. This shows how the data is being converted. Thus, the destination of the polygonal image A _k convert from polygonal image B _k, which forms a similar shape, polygonal image B _k of the conversion source in the case of FIG. 2 is a multi-conversion destination It has a resolution that is twice the height and width of the square image A _k . In this case, the polygon image A _k of the conversion destination is an isosceles triangle of the left half of the 4 × 4 square image, and the polygon image B of the conversion source is
_k can also be viewed as an isosceles triangle in the left half of an 8 × 8 square image. In this case, it isosceles triangle of another way to form a square, a polygon image A _'k and conversion source polygon image B' _k of the destination as shown in FIG. Similarly, the pixels in the source polygon image B ′ _k are converted to the destination polygon image A ′ _k
Have been converted to the pixels within. Therefore, a 4 × 4 square image is formed by the polygon images A _k and A ′ _k of the conversion destination, and an 8 × 8 square image is formed by the polygon images B _k and B ′ _{k of} the conversion source. By converting all the polygon images constituting the screen (for example, 101 in FIG. 1) into two isosceles triangular polygon images as described above, by performing a conversion process on these polygon image units, The conversion process in the polygon conversion unit 2 can be performed without lack. Although an isosceles triangular polygon image has been described as an example here, it goes without saying that any other shape may be used as long as the screen can be configured without any excess or shortage (spreading without gaps). .

FIG. 4 shows an embodiment in the case of a polygon image of a square block. In the figure, a 4 × 4 block B _k
Is converted into a 2 × 2 block A _k . The advantage of using such a square block is simplicity in hardware configuration. On the other hand, the advantage of the above-mentioned isosceles triangular polygon image is that it is efficient with respect to oblique edge components that are often included in the image. Therefore,
In this case, a polygonal image obtained by dividing a square block into two is more likely to improve the image quality of the decoded image.

Next, a second embodiment according to the present invention will be described with reference to FIG. FIG. 5 is a block diagram illustrating a configuration example of an image decoding device according to the second embodiment. The configuration of this apparatus is basically similar to the configuration of the image decoding apparatus described in the first embodiment in FIG. The difference is that it has a resolution conversion unit. The resolution converter corresponds to the upsamplers 6 and 7 in FIG. The upsamplers 6 and 7 perform upsampling at a resolution of a predetermined magnification.

Next, the operation will be described. As in the first embodiment, a decoded image 101 decoded by the reference image decoding unit 1 is obtained. Next, the decoded image 109 having the original resolution of N ₁ times is obtained by the upsampler 6 which increases the resolution by N ₁ times in the horizontal and vertical directions. The decoded image 109 is input to the polygon image conversion unit 2 according to the processing procedure already described,
In the same part, conversion in units of polygon images is performed, and the converted image 11
0 is output. This is further added to the decoded image 109 in the adder 3 to obtain the first added image 111 as the decoded image 4 (I ₁ ).

The decoded image 4 (I ₁ ) is further increased in resolution in the horizontal and vertical directions to N ₂ , and outputs a decoded image 113 having a resolution of N ₂ times the original resolution. Subsequent processing is the same as described above. Although FIG. 2 shows only the polygon conversion processing up to the second stage, it is apparent that the processing is continued if further processing is required. is there.

The present embodiment is an effective example when the decoded image is successively zoomed, and is effective, for example, when the texture frequently used in computer graphics or the like is enlarged or reduced. . The feature of the example of FIG. 5 is that the decoding is performed while changing the zoom ratio (N ₁ , N ₂ times) in the process of decoding the texture. Therefore, when no zoom is involved, these processes may be omitted.

FIG. 6 is a diagram for specifically explaining the processing in the first stage. In FIG. 6, the reference image or decoded image 1 decoded by the reference image decoding unit 1 is shown.
01 is up-sampled by N ₂ times to generate a decoded image 109, and this decoded image 109 is converted to the polygon image conversion unit 2a.
The image 110 converted in step (1) and the decoded image 109 are added by the adder 3a to obtain an added image 111.

Next, an image coding apparatus according to a third embodiment of the present invention will be described with reference to FIG. FIG. 7 is a block diagram illustrating a configuration example of an image encoding device according to the third embodiment. This apparatus is for encoding an image, and corresponds to the image decoding apparatus described in the first and second embodiments.

The image encoding apparatus shown in FIG. 7 uses, as encoding information, a downsampler 8 for downsampling an input image 117 and an address between two polygonal images having different components in the downsampled image. A polygon image encoding unit 9 to be output, an upsampler 20 to upsample a downsampled image, and an upsampled image based on encoding information output from the polygon image encoding unit 9. , A polygon image decoding unit 10 for converting a polygon image, and a reference image encoding unit 25 for encoding an original image (reference image) and outputting encoded information. 117 and polygon image decoding unit 10
And a difference determination unit 29 that determines a difference from the subtractor 11 and controls a conversion process in the polygon image encoding unit 9 by determining a difference from the subtractor 11. I have.

Next, the operation will be described. Input image 11
Reference numeral 7 denotes a downsampler 8, which downsamples to a power of 2 in the horizontal and vertical directions, for example, to a resolution of 1/2, and outputs a resolution-converted original image 118. The down-sampled image 118 is then processed by the upsampler 20 in the horizontal and vertical directions to a power of 2, for example, 2
The resolution-converted image 120 that has been upsampled to double the resolution and generated is output to the polygon image decoding unit 10. Separately, the polygon image encoding unit 9 converts the polygon image of the conversion source into the polygon image of the other conversion destination for each of all the polygon images constituting the original image 118 whose resolution has been converted. The operation is performed on all the polygon images constituting the image. At this time, when a certain polygon image is converted, an operation of selecting a partner's polygon image closest to the converted polygon image is performed. For specific examples of the polygon image encoding unit 9 and the polygon image decoding unit 10,
Each will be described later with reference to FIGS.

As a result, the polygon image encoding unit 9 outputs
Conversion information 108 of all polygon images constituting the screen
Is sent. As is clear from FIG. 2 and the like, the present encoding apparatus searches for a polygon image having a similar shape on the screen for each polygon image (image such as an isosceles triangle or a square),
The obtained positions (addresses) of the source and destination polygon images in the screen are transmitted as encoded information 108.

Here, a specific example and operation of the detailed configuration of the polygon image encoding unit 9 will be described with reference to FIG. The polygon image encoding unit shown in FIG. 8 includes a first polygon image generation unit, a second polygon image generation unit 12,
It is configured to include an approximation degree measuring unit 20, a sub-sampling unit 26, and a multiplexing unit 27.

In FIG. 8, the resolution-converted image 118 of the original image output from the downsampler 8 in FIG.
The image is sent to the first polygon image generation unit 11 and the second polygon image generation unit 12, and is divided into a plurality of polygon images forming an image. For example, the second polygon image 123 generated by the second polygon image generation unit 12 is sub-sampled by the sub-sampling unit 26, and a newly generated polygon image 133 is obtained. Further, between the polygon image 133 and the first polygon image 122 generated by the first polygon image generation unit 11, the approximation degree measurement unit 20
When the approximation degree calculation is performed and the error is larger than a predetermined threshold value, another approximation image is generated again from the approximation degree measurement unit 20 to the second polygon image generation unit 12. An instruction signal is sent out. Thereafter, the above operation is continued until the error value becomes smaller than the threshold value.

Here, the first polygonal image 122 corresponds to A _k in FIG. 2, and the sub-sampled second polygonal image 133 corresponds to B _k in FIG. In addition, in order to calculate the degree of approximation in the degree-of-approximation measuring unit 20, generally, the square error of both values may be calculated. The above operation is performed on all the first polygon images constituting the image, and the address (position on the screen) 116 of each first polygon image and the selected second polygon image And the address 119 is multiplexed by the multiplexing unit 27 and the polygon image encoded information 108 is transmitted. The above is the configuration and operation of the polygon image encoding unit 9.

FIG. 9 is a block diagram showing a configuration example of the polygon image decoding unit 10. The configuration in FIG. 9 corresponds to the configuration in which the reference image decoding unit 1 in the configuration in FIG. 1 described above is removed, and instead of the decoded image 101 from the reference image decoding unit 1 in FIG. The configuration is the same as that described with reference to FIG. 1 except that the image 120 obtained by up-sampling in step 20 is supplied to the first-stage polygonal image conversion unit 2a and the adder 3a, and a description thereof will be omitted. .

Next, the operation of the polygon image decoding unit 10 will be described. The polygon image decoding unit 10 receives the resolution converted image 120 and the polygon image conversion information 108 and outputs a decoded image 121. Here, it should be noted in the operation of the polygon image decoding unit shown in FIG. 9 that the original image 117 and the resolution-converted image 120 have the same image size, but the resolution-converted image 120 Since the image is generated by sampling and up-sampling, the information amount of the image is less than that of the original image 117. The polygon 120 is converted using the image 120 lacking this information amount as the first reference image.

Next, the operation of the reference image encoding unit 25 will be described with reference to FIG. FIG. 10 is a diagram for describing an example of the operation of the unit 25. The input original image 117 is encoded by the reference image encoding unit 25.
In FIG. 10, the information is compressed by lowering the resolution by a simple pixel thinning process. The generated reference image encoding information outputs pixel values as raw data or encodes the pixel values as encoded information 100. The operation up to this point is the operation of the part 25. As described above, the operation of restoring the reference image in FIG. 10 to obtain the decoded image 101 is an operation in the reference image decoding unit 1 in FIG. 1 or FIG. In this case, the decoded image 101 is obtained from the reference image by simple pixel interpolation or decoding and pixel interpolation. FIG.
As is clear from 0, the hardware load is very small in simple thinning and interpolation, but much of the information amount of the original image is lost.

By the above operation, the polygon image decoding unit 10
The difference between the decoded image 121 transmitted from the original image 117 and the original image 117 is calculated by the subtraction unit 28, and the error image 13
6 is subjected to threshold processing in an error determination unit 29. As a result, when the error is larger than the preset threshold value, the same unit 29 sends an instruction signal 135 to the polygon image encoding unit 9 to perform the encoding again. The subsequent operation is the same as described above.

Next, a fourth embodiment of the present invention will be described. In the fourth embodiment, as the reference image encoding unit 25 described with reference to FIG. 10, one using orthogonal transform such as DCT (discrete cosine transform) is used.

That is, FIG. 11 shows another example of the operation in the reference image encoding unit 25 described with reference to FIG. In this case, the reference image encoding unit 25 uses JPEG or MPEG
Has a DCT operation unit used in
Is the DCT coefficient. Conversely, the reference image decoding unit 1 has an inverse DCT operation unit, and decodes this to restore an image. In this case, since the distortion cannot be detected even if the DCT compresses the original image by a factor of ten, it is possible to restore a reference image of extremely high quality as compared with the example of FIG. 11 on the decoder side.

Next, a fifth embodiment according to the present invention will be described. In the fifth embodiment, a polygon image conversion unit changes the positional relationship between the position of a reference pixel to be converted and the polygon, and further, if necessary,
The positional relationship is changed according to the number of steps where the polygon image conversion unit is located.

In the embodiment already described, FIGS.
As described with reference to FIG. 4, the positional relationship between the source pixel and the destination pixel is constant. However, in the fifth embodiment, in the case of the polygon image shown in FIG.
There are two [0], [1], and [2]. Therefore, according to the number of steps of the polygon image conversion unit described in FIGS. 1 and 5, it can be set so as to go around the positions of these pixels to be referred to. This has the effect of improving image quality. Therefore, it is sufficient to determine the reference pixel position number = the number of stages% 3 (the remainder when divided by 3).

An embodiment of an image decoding device and an image encoding device according to the present invention realizes an entertainment device such as a video game machine, a video / still image codec used in the Internet, mobile communication, and the like, or the same system. It can be used for software modules and the like.

The present invention is not limited to only the above-described embodiments, and it goes without saying that various modifications can be made without departing from the spirit of the present invention.

[0051]

According to the image decoding apparatus and method of the present invention, based on an image obtained by decoding by the reference image decoding means, polygon image conversion information in the image is obtained by using the polygon image conversion information. Perform the conversion. Further, by adding this to the image before conversion, a new image is generated, and the above processing is repeated by a predetermined number of stages, whereby there is an effect that the image quality is gradually improved. The required number of stages varies depending on the image, but since the reference image generated by the reference image decoding unit is similar to the original image, a high-quality enlarged image is always obtained even if the resolution is increased and the image is enlarged. There is an effect that it can be.

Further, according to the image encoding apparatus and method according to the present invention, the image encoding apparatus includes a polygon image conversion unit and a decoding unit for performing local decoding. Moreover, since the conversion of the polygon image is repeatedly performed until the difference value between the original image and the decoded image becomes smaller than the threshold value, a high-quality encoded image can be always provided.

Further, by using a polygonal image having an isosceles triangle shape as the polygonal image, encoding can be performed with high efficiency even for oblique edge component images which frequently appear in the image. There is. On the other hand, by using a quadrangular polygonal image, there is an effect that hardware can be easily realized. Therefore, the processing can be speeded up.

Also, by keeping the positional relationship between the reference pixel of the conversion source and the polygon image of the reference destination constant,
Since the processing procedure is easy, the speed is increased. It is also easy to implement hardware.

Further, since the positional relationship between the reference pixel of the conversion source and the polygon image of the reference destination is variable, there is an effect that the image quality can be improved by using the reference from the surrounding pixels.

Further, by setting the down-sampling and up-sampling means to be a power of 2, hardware can be easily realized.

Also, in the decoding unit of the image encoding device,
By repeatedly performing the conversion of the polygon image in multiple stages, there is an effect that the image quality is gradually improved. Also,
The error determination unit takes a difference value between the original image and the decoded image, and repeatedly performs the polygon image conversion until this value becomes smaller than a predetermined threshold value, so that a high-quality encoded image is always obtained. Can be offered.

In addition, by converting pixels, that is, copying pixel values, as conversion on a pixel basis, processing can be performed at a very high speed. Therefore, hardware implementation is also easy.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an image decoding device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing an example of pixel conversion from a source polygon image to a destination polygon image.

FIG. 3 is a diagram illustrating another example of pixel conversion from a source polygon image to a destination polygon image.

FIG. 4 is a diagram showing polygon image conversion when the polygon image is a square block.

FIG. 5 is a block diagram illustrating a configuration of an image decoding device according to a second embodiment of the present invention.

FIG. 6 is a diagram for explaining an operation in a decoding device with zoom.

FIG. 7 is a block diagram illustrating a configuration of an image decoding device according to a third embodiment of the present invention.

FIG. 8 is a block diagram illustrating an example of a polygon image encoding unit.

FIG. 9 is a block diagram illustrating an example of a polygon image decoding unit.

FIG. 10 is a diagram illustrating an operation when pixel thinning is used in a reference image encoding unit and an operation when pixel interpolation is used in a reference image decoding unit.

FIG. 11 is a diagram illustrating an operation when the DCT is used in the reference image encoding unit and an operation when the inverse DCT is used in the reference image decoding unit.

FIG. 12 is a diagram related to a reference pixel position of a conversion source when a polygon image is an isosceles triangle.

FIG. 13 is a block diagram illustrating a wavelet encoding device as an example of a conventional encoding device.

FIG. 14 is a diagram illustrating a filtering operation at each level in a conventional wavelet transform.

FIG. 15 is a diagram for explaining a filtering operation at each level in the conventional wavelet inverse transform.

FIG. 16 is a diagram showing a state of wavelet division.

FIG. 17 is a diagram showing each band component image when wavelet division is performed on an actual still image.

[Explanation of Codes] 1 Reference image decoding unit, 2 Polygon image conversion unit, 3
Adder, 4 1st stage added image, 5 2nd stage added image, 6 N ₁ times up sampler, 7 N ₂ times up sampler, 8 down sampler, 9 polygon image encoding unit, 10 Polygon image decoding unit, 11
1st polygon image generator, 12 second polygon image generator, 13 wavelet transform unit, 14 transform coefficient scanning unit, 15 quantizer, 16 entropy encoder, 17 low-pass filter (analysis), 18 High-pass filter (analysis), 19
Down sampler, 20 approximation degree measuring section, 21 up sampler, 22 low-pass filter (synthesis),
23 high-pass filter (synthesis), 24 adder,
25 reference image coding unit, 26 subsampling unit, 27 multiplexing unit, 28 difference unit, 29 error determination unit

Claims (27)

[Claims] 1. An initial image generating means for inputting encoding information and generating an image serving as a source of a decoded image, and converting a polygon image into another polygon image in an image generated in each stage. And a synthesizing means for synthesizing the image before the conversion and the image after the conversion, wherein the polygon image conversion means and the synthesizing means have a multi-stage configuration. Image decoding device. 2. The polygon image conversion means performs a conversion from one polygon image to another polygon image based on polygon image encoding information for each polygon image constituting the entire screen. The image from the initial image generation means is input to the first-stage polygon image conversion means having a multi-stage configuration, and the composite image from the previous synthesis means is input to the second and subsequent polygon image conversion means, 2. The image decoding apparatus according to claim 1, wherein a synthesized image from the last-stage synthesizing means is extracted as a decoded image. 3. A resolution conversion means for converting the resolution of an image in accordance with a given enlargement / reduction ratio, wherein the resolution conversion means, the polygon image conversion means, and the synthesis means have a multi-stage configuration. 2. The image decoding apparatus according to claim 1, wherein: 4. The image from the resolution conversion means in each stage of the multi-stage configuration is sent to the image conversion unit and the synthesizing unit of the stage, and the resolution conversion unit in the first stage of the multi-stage configuration is sent from the initial image generation unit. Wherein the synthesized image from the preceding synthesizing means is input to the second and subsequent resolution converting means, and the synthesized image from the last synthesizing means is extracted as a decoded image. 3. The image decoding device according to 3. 5. The image decoding apparatus according to claim 1, wherein said polygonal image has a triangular or quadrangular shape. 6. The image decoding apparatus according to claim 1, wherein said polygon image conversion means includes means for performing pixel-by-pixel conversion of said polygon image. 7. The image decoding apparatus according to claim 6, wherein the conversion in pixel units is a transfer of pixel values. 8. The image decoding apparatus according to claim 7, wherein the transfer of the pixel value is realized by referring to an address of a memory that stores and holds an image. 9. The image decoding apparatus according to claim 6, wherein said pixel-by-pixel conversion multiplies each pixel value by a constant. 10. The image decoding apparatus according to claim 1, wherein a positional relationship between the position of the reference pixel of the conversion source and the polygon in the polygon image conversion means is always constant. 11. The image decoding apparatus according to claim 1, wherein the polygon image conversion means has a variable positional relationship between the position of the reference pixel of the conversion source and the polygon. 12. The method according to claim 11, wherein the positional relationship between the position of the reference pixel of the conversion source and the polygon changes in accordance with the number of steps where the polygon image conversion means is located.
The image decoding device according to any one of the preceding claims. 13. The method according to claim 1, wherein the synthesizing means includes means for calculating an average of an image before conversion by the polygon image conversion means and an image after conversion in the stage. The image decoding device according to any one of the preceding claims. 14. The image processing apparatus according to claim 1, wherein the synthesizing means includes means for multiplying the image before conversion by the polygon image conversion means and the converted image in the stage by a predetermined ratio and adding the multiplied image. The image decoding device according to claim 1, wherein: 15. The image decoding apparatus according to claim 1, wherein said initial image generation means includes means for generating a resolution different from the resolution of the image of the input encoded information. 16. The image decoding apparatus according to claim 1, wherein said initial image generation means comprises orthogonal inverse transform means for decoding an image. 17. An initial image generation step of inputting encoding information to generate an image serving as a base of a decoded image, and converting an image generated at each stage from a polygonal image to another polygonal image. And a synthesizing step of synthesizing the image before conversion and the image after conversion. The polygon image conversion step and the synthesizing step are configured in a multi-stage configuration. Image decoding method. 18. In the polygon image conversion step, for each polygon image constituting the entire screen, a polygon image is converted into another polygon image based on polygon image encoding information. In the first-stage polygon image conversion step of the multi-stage configuration, the image from the initial image generation step is input, and in the second and subsequent polygon image conversion steps, the composite image from the previous synthesis step is input, and the final 2. A composite image from the synthesizing step is extracted as a decoded image.
7. The image decoding method according to 7. 19. A resolution conversion step of converting the resolution of an image in accordance with a given enlargement / reduction ratio, wherein the resolution conversion step, the polygon image conversion step, and the synthesis step have a multi-stage configuration. 18. The image decoding method according to claim 17, wherein: 20. The image from the resolution conversion step in each stage of the multi-stage configuration is sent to the image conversion process and the synthesizing process of the stage, and in the first stage of the multi-stage configuration, the image is converted from the initial image generation process. 20. An image is input, a synthesized image from a preceding synthesis step is input in a second and subsequent resolution conversion steps, and a synthesized image from a final synthesis step is extracted as a decoded image. Image decoding method. 21. Down-sampling means for down-sampling an input image;
Polygon image encoding means for outputting an address between two polygon images as encoding information; up-sampling means for up-sampling the down-sampled image; and a polygon image for the up-sampled image. Based on the encoded information output from the encoding means, a decoding means for decoding a polygon image unit, and an image encoding means for encoding the input image and outputting encoded information, Image encoding device. 22. Difference means for calculating an error between the decoded image generated by the decoding means and the input image; and error determination means for determining the magnitude of the error from the difference means; Controlling the polygon image conversion means in accordance with the result of the determination in step (a) to search for a polygon image closest to a certain polygon image in the downsampled image. 22. The image encoding device according to 21. 23. The image encoding apparatus according to claim 21, wherein said polygon image decoding means includes means for performing a pixel-by-pixel conversion of the polygon image. 24. The downsampling means includes means for reducing the resolution to a power of two, and the upsampling means comprises means for increasing the resolution to a power of two. 22. The image encoding device according to 21. 25. A decoding apparatus comprising: means for performing a plurality of times of pixel-by-pixel conversion within a polygon on the basis of the conversion information output from the polygon image encoding means. 22. The image encoding device according to claim 21, wherein: 26. A down-sampling step of down-sampling an input image;
A polygon image encoding step of outputting an address between two polygon images as encoding information; an up-sampling step of up-sampling the down-sampled image; and a polygon image for the up-sampled image. Based on the encoded information output from the encoding step, a decoding step of decoding a polygon image unit, and an image encoding step of encoding the input image and outputting encoded information, Image encoding method. 27. A difference step for obtaining an error between the decoded image generated in the decoding step and the input image, and an error determination step for determining a magnitude of an error from the difference step. Controlling the polygon image conversion step according to the result of the determination in step (a) to search for a polygon image closest to a certain polygon image in the downsampled image. 27. The image encoding method according to 26.