KR20070111111A - Method of compressing and decompressing image and equipment thereof - Google Patents

Abstract

A method and an apparatus for compressing and decompressing an image are provided to process a small mount of data in a video RAM(Random Access Memory) served as a frame buffer by reducing the amount of video data to be processed and manufacture an encoder and an decoder by simple logic circuits, thereby improving video process rate. A method for compressing and decompressing an image comprises the following steps of: recognizing DN which is current pixel RGB data and DN-1' which is current pixel RGB data(30); computing MN which is a difference between DN and DN-1'(32); dividing MN into an up bit and a low bit and recognizing the up bit as a data bit when the up bit is not "0", and recognizing the down bit as a data bit when the up bit is "0"(34); and acquiring the DN as DN '' compressed by coding bits, position bits, and data bits(36).

Description

Image Compression Method and Extension Method and Apparatus {METHOD OF COMPRESSING AND DECOMPRESSING IMAGE AND EQUIPMENT THEREOF}

1 shows a schematic configuration of an image processing apparatus according to the prior art.

2 is a schematic flowchart of an image compression method and a decompression method according to the present invention.

FIG. 3 shows an embodiment in which the number of bits per pixel is 18 bits, and only one subpixel of each pixel is compressed by the first image compression method and expanded by the first image extension method.

4 illustrates an embodiment in which the RGB data of the first pixel on the top of the screen is recognized as the stretched data of the first pixel, compressed by the first image compression method, and expanded by the first image extension method.

FIG. 5 illustrates an embodiment in which the number of bits per pixel is 24 bits, and the virtual pixels are positioned adjacent to the first pixel of the upper screen, compressed by the first image compression method, and expanded by the first image extension method.

FIG. 6 illustrates an embodiment of compressing by the modified first image compression method of the embodiment of FIG. 5 and stretching by the modified first image stretching method.

FIG. 7 illustrates an embodiment in which the same RGB data as in FIG. 3 is compressed by the second image compression method and expanded by the second image extension method.

FIG. 8 illustrates an embodiment in which the same RGB data as in FIG. 4 is compressed by the second image compression method and expanded by the second image extension method.

FIG. 9 illustrates an embodiment in which the same RGB data as in FIG. 5 is compressed by the second image compression method and expanded by the second image extension method.

10 illustrates various methods of obtaining the difference values in the compression method and the expansion method according to the present invention.

FIG. 11 shows a picture compressed and decompressed by the second image compression method and the second image extension method.

FIG. 12 is a data table comparing various compensation values for compensating for loss of original data for the figure shown in FIG. 11.

13 shows a schematic configuration of an image processing apparatus according to the present invention.

The present invention relates to an image compression method and a stretching method and a device thereof, and more particularly, to a stable image compression method and a stretching method and a device thereof, while maintaining an image loss that can not be felt by human vision. It is about.

The whole world now lives in a flood of multimedia information. Multimedia information exists in various forms such as text, still picture, moving picture, animation and sound. Such multimedia information is changing from conventional analog information to digital information, and the amount of data of multimedia information to be processed is increasing with the rapid development of communication technology, the Internet, and display devices.

The most important part of the multimedia information may be referred to as image data. 1 is a schematic flowchart of processing existing image data. The video screen may be composed of RGB pixels of contrast or color of each point, and the central processing unit processes the contrast or color of the RGB pixels in a bitmap format and stores them in a storage unit such as a video RAM and displays some contents on the screen. To indicate, change the value of the corresponding position of the storage unit and output it to the panel through the output unit. For example, in order to process one true-calk still image with a 640 * 480 picture, one truecolor requires 24 bits per pixel and the entire screen is 640 * 480 * 24, i.e. about 7,372,800 bits about 0.9M. Takes up memory. Furthermore, a TV broadcast or video signal must process 30 to 60 frames per second in order to be recognized without interruption of the screen with human eyes. That is, to process a screen with a data capacity of about 0.9M at 30 to 60 frames per second, the storage space and memory bandwidth are required. Therefore, if you want to express more colors as a more natural image at higher resolution, more storage space and longer information processing time are required as the number of pixels and frames increases.

As described above, in order to efficiently handle the huge amount of data, various digital image data compression and decompression techniques have emerged.

Image compression algorithm can be largely divided into two methods. One is lossless coding, and the other is lossy coding.

The lossless coding method is a method of reproducing a high quality image by preserving original information even after compression and decompression. However, the lossless coding method has a lower compression rate than a lossy coding method. Not suitable for processing

On the other hand, the lossy coding method has a very high compression ratio, but distortion occurs during compression and decompression, so that the original image is not completely reproduced, thereby degrading the image quality.

A representative example of the lossy coding scheme is JPEG technology.

Joint Photographic Coding Experts Group (JPEG) is a method used to compress continuous images, such as photographs. Image images are not so much 'repetitive of information', so when compressed using the 'repeated information', the compression rate is lowered and a good result is not obtained. Thus, JPEG discards any information before compression, which is why JPEG loses its compression and reconstruction.

JPEG mainly compresses only the low-pass part, leaving only the high-pass part that is relatively sensitive to the human eye in the image data. It's a good thing.

The lossy compression of JPEG uses a lot of information that cannot be noticed by humans, so it discards this information and uses only actual data.

JPEG's lossy coding scheme is based on DCT. DCT is a kind of mapping technology that converts a spatial domain and a frequency domain into each other, and can decompose a natural image in a spatial domain from a low frequency component to a high frequency component in the frequency domain.

JPEG uses a combination of lossy coding (DCT + quantization) and lossless coding (DPCM, RLE, Huffman coding, and arithmetic compression) to increase the compression rate. Focus on and then quantize it. Looking at each step in more detail, in step 1, the DCT transform is used to ensure that 64 (8 * 8) pixels in the image have a low signal value, thereby reducing the overlapping signal value in image space to increase the compression rate, and in the second step, quantization Using a given quantization table containing the signal values that can be expressed in steps, an approximation of 64 values resulting from the DCT step is obtained. In step 3, the size of quantized image information is further reduced through entropy coding. .

The JPEG method is performed through several steps as described above, and there is a problem in that a complicated equation for each step and a device according thereto are required.

In addition, since the compression rate of the image by JPEG method varies widely according to the content of the image, when recording / reproducing or transmitting an image, the recording / fetching time or transmission time is very variable according to the content of the image to be compressed. There is a problem that is.

The present invention was devised to solve the above problems, eliminating the need for a large amount of storage space and a high clock frequency in processing image information by eliminating complicated equations and steps and by an efficient compression method, This eliminates the instability of buffering due to the variability of time required for playback, storage and transmission.

In order to achieve the above object, the present invention provides a method for recognizing current pixel RGB data D _N and expanded data D _N-1 ′ based on previous pixel RGB data D _N-1 ; Calculating a difference value of the D M _{N N} the D _N-1 'for; If the M _N is positive or zero, the sign bit is recognized as 0. If the _{N N} is negative, the sign bit is recognized as 1, and if the M _N is divided into an upper bit and a lower bit, the position bit is set to 1 if the upper bit is not 0. Recognizing the upper bit as a data bit, and if the upper bit is 0, recognizing a position bit as 0 and recognizing the lower bit as a data bit; And converting the D _N into the sign bit, the position bit, and the data bit to obtain compressed data D _N ″. Where N = 1, 2, 3, ... n, ... where n is a natural number

The present invention also provides a method for recognizing compressed data D _N ″ for RGB data of a current pixel and extended data D _N-1 ′ based on RGB data D _N-1 of a previous pixel; Calculating a sign bit, a position bit, and a data bit from the compressed data D _N ″; If the position bit is 1, the upper portion of the data bit the M _N upper bits and the M _N lower bits is recognized as 0, and if the position bit is 0 and the lower bits of the data bit M _N a M _N of the Recognizing a bit as 0, recognizing the M _N as a positive number if the sign bit is 0, and recognizing the M _N as a negative value to restore the difference value M _N ; And adding M _N to D _N-1 ′ to obtain extended data D _N ′.

And recognizing the decompression data D _N-1 ′ based on the current pixel RGB data D _N and the previous pixel RGB data D _N-1 ; Calculating a difference value of the D M _{N N} the D _N-1 'for; If the absolute value of M _N is greater than the maximum value that H bits (where H represents half of the number of bits occupied by one subpixel RGB data of the one pixel) can be recognized as a position bit 1 The data bit is recognized as a value from the upper side of the D _N to the H + 1 bit, and if the absolute value of the M _N is less than or equal to the maximum value that the H bit can have, the position bit is recognized as 0. Recognizing a data bit as a value from a lower level of M _N to a H + 1 bit; And converting the D _N into the position bits and the data bits to obtain compressed data D _N ″.

Recognizing the compressed data D _N ″ for the RGB data of the current pixel and the decompressed data D _N-1 ′ based on the RGB data D _N-1 of the previous pixel; Calculating position bits and data bits from the compressed data D _N ″; And when the position bit is 1, the upper part of D _N 'is recognized as a value of the data bit from H + 1 bit and the lower value of D _N ' is the maximum value that H-2 bit can have from H-1 bit. If the position bit is 0, the first bit of the data bit is recognized as a sign bit and the value added by the remaining bit and D _N-1 'is recognized as D _N ', thereby compressing the compressed data D. obtaining the _N '' stretched data D _N for " The image stretching method comprising the step.

Furthermore, the present invention may include obtaining compressed data for RGB data of each pixel sequentially for all pixels of the screen.

Preferably, the image compression method may include the step of recognizing the RGB data D ₁ of the first pixel on the top of the screen as the extended data D ₁ ′.

And more preferably characterized in that it comprises a step of recognizing in any RGB data D virtual pixel of having a _zero neighborhood located at the first pixel at the top of the screen and extending the D ₀ of the virtual pixel data D ₀ ' It can be an image compression method.

Furthermore, the present invention may include obtaining the decompressed data for the compressed data of each pixel RGB data sequentially for all the pixels of the screen.

In another aspect, the present invention in the calculated between the current pixel RGB data D _N and a previous pixel RGB data decompressed data based on the _{D N-1 D N-1} ' difference M _N, said calculated difference value M _N An encoder that compresses the current pixel RGB data D _N based on the result to obtain compressed data D _N ″; A frame buffer for buffering the compressed data D _N ″ calculated by the encoder; And a decoder for decompressing the compressed data D _N ″ buffered by the frame buffer to obtain the decompressed data D _N ′.

Preferably, the encoder is an image processing apparatus characterized in that it receives the extended data D _N-1 ′ based on the previous pixel RGB data D _N-1 from the decoder.

More preferably, the encoder can compress RGB data by one of two image compression methods according to the present invention, and the decoder can decompress the compressed data by one of two image extension methods according to the present invention.

The present invention may also be an image processing apparatus including an interface for outputting an RGB data signal of each pixel of an input image to an encoder and an output unit for outputting D _N ′ extended by the decoder.

Hereinafter, an image compression method, an extension method, and an apparatus thereof according to the present invention will be described in detail with reference to the accompanying drawings.

2 is a schematic flowchart of an image compression method and a decompression method according to the present invention.

Referring to (a) of FIG. 2, a method of compressing an image according to the present invention will be described. When an RGB signal of an image is input, the image compression method is extended based on the RGB data D _N of the current pixel and the RGB data D _N-1 of the previous pixel. 'going through the step (30) for recognizing, D _N-1 for the _N D, the data D _N-1 subjected to step 32 to obtain the difference value of M _N.

Subsequently, one of the image compression methods according to the present invention passes through step 34 of generating a sign bit, a position bit, and a data bit based on the difference value M _N , wherein D _N is represented by the sign bit, the position bit, and the data bit. A compressed video signal is generated through step 36 of obtaining compressed data D _N ″ as data bits. Hereinafter, the image compression method will be referred to as a first image compression method.

Another one of the image compression methods according to the present invention is the step (34) of generating position bits and data bits based on the difference value M _N , and the data D compressed with D _N as the position bits and the data bits. _Through step 36 of obtaining _N '', a compressed video signal is generated. Hereinafter, the image compression method will be referred to as a second image compression method.

In addition, referring to FIG. 2B, the decompression method according to the present invention will be described. When the compressed data is input, the decompressed data D _N-1 ′ of the previous pixel and the compressed data D _N ′ of the current pixel are input. 'going through the step 40 to recognize, one of the image height the process according to the invention the D _N' going through the step 42 to calculate a "sign bit, where the bit and data bits, the sign bit, the positional bits And reconstructing the difference value M _N based on the data bits (44), and adding the difference value M _N to the extended data D _N-1 ′ of the previous pixel to obtain the expanded D _N ′ (46). The video signal extended through) is obtained. Hereinafter, the image decompression method will be referred to as a first image decompression method.

Another one of the image height the process according to the invention the height data of the D _N ', the position bits and data bits go through the step 42 of calculating a D _N height to compensate for the data bits' Obtain or previous pixel The extended video signal is obtained through the step 46 of adding the data bits to D _N-1 'to obtain the expanded D _N '. Hereinafter, the image decompression method will be referred to as a second image decompression method.

3 to 6, the first image compression method and the first image extension method according to the present invention will be described in detail with reference to an embodiment.

FIG. 3 illustrates an embodiment in which only one subpixel of each pixel is compressed by the first image compression method according to the present invention and expanded by the first image extension method as part of a screen having 18 bits per pixel.

When the number of bits per pixel is 18 bits, each subpixel of R, G, and B occupies 6 bits, and one subpixel of FIG. 3 is 6 bits, so the maximum value of one subpixel data is 2 ⁶ -1, that is, 63 do.

A first image compression method and a first image extension method will be described with reference to FIG. 3.

If the RGB data D _N 210 of any N th pixel is 55 and the extended data D _N-1 ′ 202 of the N-1 th pixel, which is before the N th pixel, is 50, for D _N 210. When the difference value M _N of D _N-1 ′ 202 is calculated, the difference value M _N becomes 5.

In the first image compression method, a sign bit is recognized as 0 if the difference value is positive or zero, and a sign bit is recognized as 1 if the difference value is negative, and the difference value is binary converted into the same number of bits as the number of bits occupied by the RGB data. To separate the upper and lower bits.

In this case, since the difference value M _N is positive as 5, the sign bit is recognized as 0. The binary conversion of the difference value M _N is 000101, the upper bit is 000, and the lower bit is separated by 101.

If the upper bit is not 0, the position bit is 1 and the upper bit is recognized as the data bit. If the upper bit is 0, the position bit is 0 and the lower bit is recognized as the data bit. In this case, since the upper bit is 0, the position bit is 0 and the lower bit 101 is recognized as a data bit.

Eventually it is recognized as sign bit 0, position bit 0, and data bit 101 so that D _N 210 is compressed to 5 bits of 0-0-101 (211).

A first image decompression method for decompressing the compressed data D _N ″ 211 of the N th pixel will now be described. Recognizing the compressed data D _N '' 211 and the stretched data D _N-1 ′ 202 of the immediately preceding pixel, D _N '' 211 is 0-0-101 and D _N-1 '( 202 has a value of 50.

If D _N '0-0-101 from the' 211 calculates the sign bit, bit position, and a data bit, the sign bit is 0, the bit position is 0 and the data bit is calculated as 101.

The first image decompression method recognizes the data bit as an upper bit of the difference value M _N and the lower bit of M _N as 0 when the position bit is 1, and recognizes the data bit when the position bit is 0. The lower bit of the difference value M _N and the upper bit of the M _N are recognized as zero. Therefore, D _N, the sign bit of the '211 is zero because the data bits 101 and the lower bits of the difference M _N upper bits of the M _N is recognized to zero. Therefore, M _N becomes 000101 to be 5.

When the difference value M _N is restored, the difference value M _N is added to the extended data D _{N -1} ′ 202 of the previous N−1th pixel to obtain the expanded data D _N ′ 212. 50 of the stretched data D _N-1 ′ 202 of the pixel is added to 5 of M _N to 55, and the extended data D _N ′ 212 becomes 55.

Further, when the compressed data and the decompressed data of the RGB data of each pixel are sequentially obtained for all the pixels of FIG. 3, D _{N + 1} 220 of the N + 1 th pixel is 55 and the previous pixel is obtained. If the extended data D _N ′ 212 of the N th pixel is 55, and the difference value M _N of D _N ′ 212 with respect to D _{N + 1} 220 is calculated, the difference value M _{N + is} 55-55. ₁ becomes 0.

Since the difference value M _{N + 1} is 0, the position bit becomes 0. When the difference value M _{N + 1} is binary-converted and separated into an upper bit and a lower bit, it becomes 000 and 000, respectively. Since the upper bit is 0, the data bit is the lower bit 000.

Therefore, the compressed data D _{N + 1} ″ 221 for D _{N + 1} 220 becomes 0-0-000.

To expand the compressed data D _{N + 1} '' 221 by the first image stretching method, 0-0-000 of D _{N + 1} '' 221 and the extended data D _N ′ 212 of the previous pixel. of recognizing the 55, D _{N + 1} ", the sign bit from the 0-0-000 221 0, bit position 0, and the output data bit 000 and, D _{N + 1,} the position of the '221 Since the bit is zero, the data bit 000 becomes the lower bit of the difference value M _{N + 1} and the upper bit of the M _{N + 1} is recognized as zero. Therefore, M _{N + 1} becomes 000000 and becomes 0. If 55 of the extended data D _N '212 of the previous pixel is added to 000000 of M _{N + 1,} the extended data D _{N + 1} ' ( 222 becomes 55.

By compressing 63 of the RGB data D _{N + 2} 230 of the N + 2 th pixel by the above-described method, D _{N + 2} '' 231 becomes 0-1-001, and the compressed data D _{N + 2} ' Increasing '(231) D _{N + 2} ' (232)

63.

N + 3 beonjjae gritty pixel compress D _{N + 3} (240) in, N + 3 th pixel of D _{N + 3} (240) is 47 and the height data of the (N + 2) th pixel is the previous pixel D _{N + 2} ' If (232) is 63, and the difference value M _{N + 3} of D _{N + 2} '(232) with respect to D _{N + 3} 240 is calculated, the difference value M _{N + 3} becomes -16. .

With the use of two's complement to represent a negative number in binary, the difference value M _{N + 3} is negative, because the sign bit is a 1, the difference value M _{N + 3} are negative, so to convert binary numbers 2 If you take the complement of, it becomes 1110000. If you divide the upper bit and lower bit except the sign bit, it becomes 110 and 000, respectively. Since the upper bit is not zero, the data bit becomes the upper bit 110.

Accordingly, the compressed data D _{N + 3} ″ 241 for the D _{N + 3} 240 of the N + 3 th pixel becomes 1-1-110.

To decompress the compressed data D _{N + 3} '' 241 by the first image stretching method, 1-1-110 of D _{N + 3} '' 241 and the decompressed data D _{N + 2} 'of the previous pixel. 232), calculates the sign bit 1, the position bit 1, and the data bit 110 from 1-1-110 of D _{N + 3} '' 241, and calculates the value of D _{N + 3} '' 241. Since the position bit is 1, the data bit 110 becomes the upper bit of the difference value M _{N + 3} and the lower bit of the M _{N + 3} is recognized as 0. Therefore, M _{N + 3} is 110000, but since the sign bit is 1, it becomes a two's complement of negative number, and 110000 is 111000, which is converted to decimal, so that M _{N + 3} becomes -16.

When 63 of the extended data D _{N + 2} ′ 232 of the previous pixel is added to −16 of M _{N + 3} , 47 becomes 47 and the expanded data D _{N + 3} ′ 242 becomes 47.

In order to compress the RGB data according to the present invention, since the extended data of the previous pixel is required, in FIG. 4, the RGB data D ₁ of the first pixel at the top of the screen is recognized as the extended data D ₁ ′ of the _first pixel. Compresses and decompresses the RGB data D ₂ at the second pixel.

2, 18 is an elongated data of the first pixel of the previous pixel D ₁ D ₂ (320) of the second pixel (312) as a 47, D ₁ to D ₂ (320), the difference value of 312 M ₂ The difference value M ₂ is -29.

Since the difference value M ₂ is negative, the sign bit is 1, and the difference value M ₂ becomes 1100011 when taking a two's complement to perform binary conversion, except for the first bit indicated by the negative representation of the two's complement. And the upper and lower bits are 100 and 011, respectively. Since the upper bit is not zero, the data bit becomes the upper bit 100.

Therefore, the compressed data D ₂ ″ 321 of the second pixel D ₂ 320 becomes 1-1-100.

To expand the compressed data D ₂ '' 321 by the first image stretching method, 1-1-100 of D ₂ '' 321 and 47 of the extended data D ₁ ′ 312 of the previous pixel are recognized. Calculates the sign bit 1, the position bit 1, and the data bit 100 from 1-1-100 of D ₂ '' 321, and the data bit 100 because the position bit of D ₂ '' 321 is 1, This bit becomes the upper bit of the difference value M ₂ and the lower bit of the M ₂ is recognized as zero. Thus, the M ₂ are termed the 100,000 because the sign bit is 1 is the two's complement of a negative number is 100000 converting it to a decimal number to the M ₂ 1100000 is that -32.

47 of the stretched data D ₁ ′ 312 of the previous pixel is added to −32 of M ₂ to 15, resulting in 15 stretched data D ₂ ′ 322.

For each of the remaining pixels, the compression is performed by the first image compression method and the first image decompression method, so that the deduction can be inferred from the above description, and 63 of D ₃ 330 and 63 of D ₄ 340 are the first image. Compression method compresses to 0-1-110 of D ₃ '' (331) and 0-0-000 of D ₄ '' (341) respectively, and 0-1-110 of D ₃ '' (331). And 0-0-000 of D ₄ ″ 341 are extended to 63 of D ₃ ′ 332 and 63 of D ₄ ′ 342 when extended by the first image stretching method.

Furthermore, FIG. 5 is a part of a screen in which the number of bits per pixel is 24 bits and one subpixel is 8 bits. Only one subpixel of each pixel is compressed by the first image compression method and expanded by the first image extension method according to the present invention. An embodiment is shown.

In FIG. 5, the virtual pixels are positioned adjacent to the first pixel of the screen and the RGB data of the virtual pixel is set to D ₀ , and the RGB data of each pixel is sequentially compressed using the first image compression method of the present invention. Stretch by the image stretching method.

The virtual pixel D ₀ 400 may be arbitrarily set and may be fixed to an appropriate value according to a situation. For the D ₀ (400) of the virtual pixel is a the D ₀ (400) according to was set to the highest value 255 of the RGB data, compressing the first D ₁ (410) of the second pixel, and height height data D ₀ '( 402).

The D ₁ 410 of the top first pixel of the screen is 251 and the stretched data D ₀ '402 of the imaginary pixel is 255, and the difference of D ₀ ' 402 for D ₁ 410 is M ₁ The result is -4.

Since the difference value M ₁ is negative, the sign bit is 1, and when the binary conversion is performed, the sign bit is 111111100. When the binary bit is separated into the upper bit and the lower bit except for the sign bit 1 of the first bit, the number is 1111 and 1100. The data bit is the upper bit 1111.

Therefore, the compressed data D ₂ ″ 411 of the first pixel D ₁ 410 becomes 1-1-1111.

To decompress the compressed data D ₁ '' 411 by the first image decompression method, 1-1-1111 of D ₁ '' 411 and 255 of the decompressed data D ₀ '' 402 of the previous pixel are recognized. And a sign bit 1, a position bit 1, and a data bit 1111 from 1-1-1111 of D ₁ '' 411, and since the position bit is 1, the data bit 1111 is the upper bit of the difference value M ₁ . And the lower bit of M ₁ is recognized as zero. Therefore, M ₁ is 11110000, but since the sign bit is 1, 111110000 is a two's complement of a negative number. When M ₁ is converted to a decimal number, M ₁ becomes -16.

Adding -16 of M _{1 to} 255 of the decompressed data D ₀ ′ 402 of the previous pixel results in 239, resulting in 239 of decompressed data D ₁ 412.

As the first image compression method for the remaining pixels and the first image expansion method, the decompression may be inferred as described above, and thus the description thereof is omitted.

It should be noted that the compression and decompression according to the present invention reduces the error of the error while passing through several neighboring pixels even if an error of the original data occurs, and thus the decompressed data follows the original data. .

In the above embodiments, one error occurs between the D ₂ 251 239 of the D ₂ 'of the D ₁ and D ₁ 251 251 going to the error is reduced. This is an effect that occurs because the difference between the stretched data of the previous pixel and the RGB data of the current pixel is the difference value.

In particular, in the case of a video screen, the RGB data of neighboring pixels is gradually changed rather than a sudden change in RGB data, so that the effect of following the original data of the stretched data according to the present invention does not cause a human visual error. . This effect is not limited to the first image compression method and the first image decompression method but also to the second image compression method and the second image decompression method which will be described later.

In addition, in compressing RGB data by the first compression method according to the present invention, a method of adding two's complement may be used as a method of obtaining the difference value.

Referring to FIG. 5, a process of obtaining a difference value by subtracting a two's complement is explained. D ₀ 400 of a virtual pixel is a maximum value 255 of RGB data, and D ₀ 400 is expanded data D ₀ '( If it is recognized as a binary number and converted to a binary number, it becomes 11111111, and the D ₁ 410 of the upper first pixel of the screen is 251, and if it is converted into a binary number, it becomes 11111011.

D ₁ (410) the binary value of D ₀ when the binary value of 402, 402, the difference value D ₀ to obtain the M ₁ of the "taking a 2's complement of being 100000001, D ₁ (410) to 11,111,011 Add the two's complement value 100000001 of D ₁ 510 to 111111100.

Here, the first bit of 111111100 is a two's complementary sign bit with respect to a negative number, where 1 represents a negative number and 0 represents a positive number. Thus, since the sign bit is 1, the difference value M ₁ is negative. If the 11111100 of the difference value M ₁ excluding the sign bit is divided into 4 upper bits and 4 lower bits, the upper and lower bits may be represented as 1111 and 1100, respectively, and the position bit is 1 because the upper bit is not 0. The upper bit 1111 is recognized as a data bit.

Thus, D ₁ 410 is compressed into sign bit 1, position bit 1 and data bit 1111 into 1-1-1111 of D ₁ ″ 411.

As a result, the same result is obtained even if the difference value is obtained by subtracting the two's complement.

Furthermore, FIG. 6 illustrates an embodiment in which the embodiment of FIG. 5 is modified from the first image compression method and the first image expansion method according to the present invention, and is positively compressed or decompressed when the difference value is negative.

The D ₁ 410 of the top first pixel of the screen is 251 and the stretched data D ₀ '(404) of the imaginary pixel is 255, and the difference of D ₀ ' (404) for D ₁ 410 is M ₁ The result is -4.

Since the difference value M ₁ is a sign bit to know the sign of the difference value M ₁ , for convenience of calculation, when the M ₁ is positively calculated to calculate the position bit and the data bit, the sign bit of −4 of M ₁ is 1; If -4 is positive, it becomes 4, and if it is converted to binary, it becomes 00000100. When the upper bit and the lower bit are separated into 0000 and 0100, respectively, since the upper bit is 0, the position bit becomes 0 and the data bit becomes the lower bit 0100.

Therefore, the compressed data D ₂ ″ 413 for the D ₁ 410 of the first pixel is 1-0-0100.

To decompress the compressed data D ₁ '' 413 by the first image decompression method, 1-0-0100 of D ₁ '' 413 and 255 of the decompressed data D ₀ '' 404 of the previous pixel are recognized. Calculates the sign bit 1, the position bit 0, and the data bit 0100 from 1-0-0100 of D ₁ '' 413, and since the position bit is 0, the data bit 0100 becomes the lower bit of M ₁ and The upper bit of M ₁ is recognized as zero. Therefore, M ₁ becomes 00000100, which is 4 when converted to decimal. Since the sign bit represents a negative number as 1, the M ₁ becomes negative and becomes -4.

When -4 of M ₁ is added to 255 of the extended data D ₀ '404 of the previous pixel, 251 becomes 251, and the expanded data D ₁ 414 becomes 251.

Compression and decompression for the remaining pixels are omitted since they can be inferred from the above description.

In this way, when the difference value is positively compressed and expanded, it is not necessary to use a two's complement calculation to express negative numbers.

As described above, when the image is processed by the first image compression method and the first image expansion method according to the present invention, if one pixel is 18 bits in a simple step, one subpixel is compressed into 5 bits and one pixel is compressed. If one pixel is compressed to 15 bits, and one pixel is 24 bits, one subpixel is compressed to 6 bits, and one pixel is compressed to 18 bits. Therefore, the amount of image data to be processed is reduced, and a smaller amount of data is processed in the video RAM serving as a frame buffer, thereby improving image processing speed.

7 to 9 are examples of a second image compression method and a second image expansion method according to the present invention, in which RGB data of each pixel is identical to those of the first image compression method and the first image extension method according to the present invention. 9, a description will be given of the second image compression method and the second image expansion method without any overlapping portions.

'And 206 is 50, D _N-1 to D _N (210), FIG. 7, the RGB data D _N (210) of the N-th pixel is 55 and the height data D _N-1 of the previous pixel 206 The difference value M _N is 5, and if it is converted to binary, it is 000101.

In the second image compression method of the present invention, half of the number of bits occupied by RGB data of one subpixel of one pixel is represented by H. In this embodiment, one pixel is 18 bits and RGB data of one subpixel is Since the number of bits occupied is 6, H becomes 3.

Also, in the second image compression method, if the absolute value of the difference value M _N is larger than the maximum value of the H bit, the position bit is 1 and the data bit is recognized as the value from the top of D _N to the H + 1 bit. If the absolute value of the M _N is less than or equal to the maximum value that the H bit can have, the position bit is set to 0, and the upper bits of the M _N and H + 1 bits are recognized as data bits.

In the case of the embodiment, since M _N is 5, which is the maximum value that 3 bits H bits can have, that is, less than 7, position bits are 0 and data bits are recognized as values from the lower part of M _N to 4 bits. If 5 of M _N is converted to binary, the value is 000101, and the value from the lower part of M _N to 4 bits is 0101. Therefore, when compressing a D _N (210) in a second compression process according to the invention are compressed to the position 0 and the bit of data bits _{0101, D N '' (215} ) is a 0-0101.

The second image decompression method for decompressing the compressed data D _N ″ 215 of the N th pixel will now be described. Recognizing the compressed data D _N '' 215 and the stretched data D _N-1 '' 206 of the immediately preceding pixel, D _N '' 215 is 0-0101 and D _N-1 '' 206. Has a value of 50.

Computing the position bits and the data bits from 0-0101 of D _N '' 215 yields the position bits as 0 and the data bits as 0101.

In the second image decompression method according to the present invention, when the position bit is 1, the upper part of the data bit is recognized as the value of the data bit from the upper part of D _N 'to the H + 1 bit, and the lower part of the H to the H-1 bit from the lower part of D _N '. If the position bit is 0, the first bit of the data bit is recognized as a sign bit, and the value of the remaining bits plus D _N-1 'is added to D _N. To be recognized.

In the second image decompression method, when the position bit is 1, the reason for recognizing the lower value from D _N 'to the H-1 bit as the maximum value that the H-2 bit can have is the second image compression method according to the present invention. Since the loss of the original data occurs when the image is compressed and expanded by the second image stretching method, the loss of the original data is compensated for, and will be described in detail later with reference to FIG. 11.

In the case of the above embodiment, since the position bit is 0, 0, which is the first bit of the data bit of D _N '' 215, is recognized as a sign bit, and the first bit in the data bit at 50 of D _N-1 '. the remaining 101 a D _N value is restored, except that deohaejun 'is therefore D _N "is 50 + 5, i.e., 55.

Furthermore, if all pixels on the screen are sequentially compressed using the second compression method for the RGB data of each pixel and expanded using the second image stretching method, D _{N + 1} 220 is 55 and D _N '(216) is 55, and the difference value of D _N '(216) with respect to D _{N + 1} (220) M _{N + 1} is 0, and M _{N + 1} is smaller than the maximum value 3 bits can have, so the position bit is It is 0 and the data bit is recognized as a value from the lower 4 bits of M _{N + 1} to 0000. Thus, compressing D _{N + 1} 220 with the second compression method according to the present invention compresses to position bit 0 and data bit 0000, such that D _{N + 1} ″ 225 becomes 0-0000.

In addition, D _{N + 1} '' gritty to the height 225 to the second image height _{method, D N + 1 '' (} 225) is 0-0000 and D _N '(216) 55 a, D _{N + 1'} The position bit 0 and the data bit 0000 are calculated from '(225), and since the position bit is 0, the first bit 0 of the data bit represents a positive number and the remaining 000 is added to 55 of the D _N ' 216 to be 55. Therefore, D _{N + 1} ", the D _{N + 1} to the height 225, 226 is 55.

Now let's explain the compression of D _{N + 2} (230) of the N + 2th pixel: D _{N + 2} (230) is 63, D _{N + 1} '(226) is 55, and D _{N + 2} (230) For D _{N + 1} '(226), the difference value M _{N + 2} is 8 and M _{N + 2} is greater than the maximum value 7 that 3 bits can have, so the position bit is 1 and the data bit is D _{N + 2} (230) is recognized as a value from 4 bits up to the binary value 111111 and becomes 1111. Therefore, compressing D _{N + 2} 230 with the second compression method according to the present invention compresses the position bit 1 and the data bit 1111, so that D _{N + 2} ″ 235 becomes 1-1111.

In addition, if D _{N + 2} '' 235 is extended by the second image stretching method, D _{N + 2} '' 235 is 1-1111, D _{N + 1} ′ (226) is 55, and D _{N +} Calculates position bit 1 and data bit 1111 from ₂ " (235), and since the position bit is 1, up to 4 bits which are H + 1 bits from the top of D _{N + 2} 'are recognized as data bit 1111 and D _{N + 2} Up to 2 bits, which are H-1 bits from the lower part of ', are recognized as 1, which is the maximum value that 1 bit, which is H-2 bits, can have. As a result, D _{N + 2} 'becomes 1111 from the upper 4 bits, becomes 01111 from the lower 2 bits, becomes 111101, and converts it to decimal to 61. Thus, the stretched D _{N + 2} ′ 236 for the compressed D _{N + 2} ″ 235 becomes 61.

In FIG. 8, the RGB data D ₁ of the first pixel at the top of the screen is recognized as the extended data D ₁ ′ of the first pixel, and the RGB data D ₂ is compressed and decompressed at the second pixel, which is the next pixel.

2, 18 is an elongated data of the first pixel of the previous pixel D ₁ D ₂ (320) of the second pixel (316) as a 47, D ₁ to D ₂ (320), the difference value of 316 M ₂ The difference value M ₂ is -29.

In this embodiment, H is 3 because one pixel is 18 bits and the number of bits occupied by RGB data of one subpixel is six.

Since the absolute value of M ₂ is 29, which is greater than the maximum value 7 that 3 bits, which are H bits, may have a position bit of 1 and a data bit is recognized as a value from 4 to 4 bits above D ₂ 320, D _When 18 of ₂ (320) is converted to a binary number, 010010 is obtained and values from the upper 4 bits of 010010 to 0100 are 0100. Thus, compressing D ₂ 320 with the second compression method according to the present invention compresses to position bit 1 and data bit 0100, resulting in D ₂ ″ 325 being 1-0100.

This time recognizes the height data D ₁ '(316) of a "will be described with respect to a second image height how to extend the 325, the compressed data D _2' compressed data D ₂ '' (325) and the immediately preceding pixel D ₂ '' 325 is 1-0100 and D ₁ '' 316 has a value of 47.

If the position bit and the data bit are calculated from 1-0100 of D ₂ '' 325, the position bit is 1 and the data bit is 0100. Since the position bit is 1, H + 1 bit is higher than D ₂ ′. Up to 4 bits are recognized as the data bit 0100 and up to ₂ bits, which are H-1 bits from the lower part of D ₂ ', are recognized as 1, which is the maximum value that 1 bit, which is H-2 bits, can have. As a result, D ₂ 'becomes 0100 from the upper 4 bits and 01 from the lower 2 bits to 010001, which is 17 when converted to decimal. Thus, the stretched D ₂ '' 326 for the compressed D ₂ '' 325 is 17.

Further if in sequence for the remaining pixels on the screen compressed in the second image coding method for the RGB data of each pixel, and extending in a second image height method, the process is omitted because it can be inferred by the above-described information, D ₃ '' (335) becomes 1-1111 and the elongated D ₃ '(336) becomes 61, D ₄ ''(345) becomes 0-1110 and the elongated D ₄ ' (346) equals 63 do.

In FIG. 9, the virtual pixels are positioned adjacent to the first pixel of the screen and the RGB data of the virtual pixel is set as D ₀ , and the RGB data of each pixel is sequentially compressed by the second image compression method of the present invention. Stretch by the image stretching method.

In the above embodiment, one pixel is 24 bits, and the number of bits occupied by RGB data of one subpixel is 8, and H is 4.

The D ₁ 410 of the top first pixel of the screen is 251 and 255 of D ₀ 400 of the imaginary pixel is recognized as extended data D ₀ '(406), and D ₀ ' for D ₁ 410 The difference value M ₁ of (406) is -4.

An absolute value of the M ₁ is less than the maximum 15 that can be the H bit of 4 bits have a 4-position bit is 0, the data bit is so recognized from the lower of said M ₁ to a value of up to 5 bits of H + 1 When the -4 of M ₁ is converted into a binary number, it is 111111100, and the value from the lower part of the 111111100 to 5 bits is 11100. Thus, compressing D ₁ 410 with the second compression method according to the present invention compresses to position bit 0 and data bit 11100, such that D ₁ ″ 415 becomes 0-11100.

This time recognizes the height data D ₀ '(406) of a "will be described with respect to a second image height method for extending the 415, the compressed data D _1' of the compressed data D ₁ '' (415) and the next pixel D ₁ '' 415 is 0-11100 and D ₀ '' 406 has a value of 255.

Computing the position bits and the data bits from 0-11100 of D ₁ '' 415 results in the position bits being 0 and the data bits being 11100. Since the position bits are 0, the first bit of the data bits, 1, is a sign. Converting 11100, which represents negative twos in bits and two's complement of negatives, to -4 results in -4 Accordingly, the value obtained by adding -4 to D ₀ ′ is recognized as D ₁ ′. Eventually the stretched D ₁ '(416) is 251, plus -4 to 255.

Further, if the remaining pixels of the screen are sequentially compressed with the second compression method for the RGB data of each pixel and then expanded with the second image extension method, the process can be inferred from the above description, and thus, D ₂ '' ( 425 becomes 1-00110 and the elongated D ₂ ′ (426) becomes 51, and the D ₃ '' (435) becomes 1-1111 and the elongated D ₃ ′ (436) becomes 247. .

10 illustrates various methods of order of obtaining the difference values.

In Figure 10 (a) is a method for obtaining the difference value along each row based on the first pixel of the top screen, Figure 10 (b) is the difference value in a zigzag form based on the first pixel of the top screen In FIG. 10, (c) illustrates a method of obtaining the difference value based on a virtual pixel. Further, the order of obtaining the difference value of the present invention may be modified in various ways without being limited to the above embodiments.

11 shows a picture compressed and decompressed by the second image compression method and the second image decompression method, where original represents the original image screen and convert converts the original image screen by the compression method and the decompression method according to the present invention. It is a screen stretched after one. As shown in FIG. 11, the present invention has a low error that is difficult to discern with the human eye, and compresses and stretches the video signal in a simple step, thereby reducing the amount of video data to be processed and increasing the processing speed. .

FIG. 12 is a data table comparing various compensation values for compensating for loss of original data for a picture compressed and expanded by the second image compression method and the second image extension method shown in FIG. 11.

The picture of FIG. 11 is represented by 18 bits per pixel, and one subpixel has 6 bits. The above figures were added with compensation values of 00, 01, 10, and 11 to compensate for the loss of data during compression and compression in accordance with the present invention. According to the present invention, when the pictures are decompressed by the second image decompression method, if the position bit is 1, H + 1, i.e., up to 4 bits, has the original data value D _N from the upper part of the data D _N 'restored. Since -1 bit, or 2 bits, is lost, it is to find the optimal compensation value.

As shown in the data table shown, (a) at sunset, a minimum error occurs when 10 compensation is added, while the remaining (b) winter, (c) training, and (d) blue hills add compensation 01. If you do, the minimum error occurs.

Therefore, in the present invention, the maximum value that the H-2 bits can have with respect to the lost H-1 bits is set as a compensation value in order to be applied to all cases with a certain uniformity. In the case of the above figure, 01, which is the maximum value that H-2 bits, that is, one bit can have, is the compensation value according to the present invention.

Therefore, as mentioned briefly above, in the second image decompression method, when the position bit is 1, the second image according to the present invention is recognized as the maximum value that the H-2 bit can have from the lower part of D _N 'to the H-1 bit. The expansion method compensates for the loss of the original data at the time of stretching.

As described above, the second image compression method and the second image decompression method according to the present invention compensate for the loss of data that may occur during compression and decompression, and reduce the amount of image data to be processed in a simple step. The video RAM, which acts as a frame buffer, processes less data, which speeds up image processing.

13 shows a schematic configuration of an image processing apparatus according to the present invention.

The main components of the image processing apparatus of the present invention are an encoder 614, a storage 616, and a decoder 618.

Encoder 614 at the current pixel RGB data D _N on the basis of the current difference between the value of the data D _N-1 'height based on the RGB data D _N-1 of a previous pixel to the RGB data of the pixels D _N M _N compressed to generate a compressed data D _N ''.

In the encoder 614, D _N may be compressed data D _N ″ compressed by the first image compression method or the second image compression method according to the present invention.

In the encoder 614, the compressed data D _N ″ is buffered in the storage unit 616 serving as a frame buffer, and the buffered D _N ″ is expanded in the decoder 618.

Decoder 618. The 'is extended in the first method or the second image height image height method in accordance with the present invention D is _{N, N} compressed the D' may be.

In addition, the D _N ′ extended by the decoder 618 is sequentially transferred to the encoder 614 to calculate a difference value M _{N + 1} from the RGB data D _{N + 1} of neighboring pixels.

The image processing apparatus according to the invention is for outputting RGB data signal of each pixel of the input image wherein the interface (612) and an elongated data D _N 'for delivery to the encoder 614 to the display panel output ( 620).

The image processing apparatus according to the present invention compresses the original RGB data by the simple compression method according to the present invention, stores the data in a frame buffer, and buffers the data, thereby reducing the capacity of the video RAM and improving image processing time.

The image compression and decompression method and apparatus according to the present invention can be modified and applied in various forms within the scope of the technical idea of the present invention and are not limited to the above embodiments. In addition, the embodiments and drawings are merely for the purpose of describing the contents of the invention in detail, and are not intended to limit the scope of the technical idea of the invention, the present invention described above is common knowledge in the technical field to which the present invention belongs As those skilled in the art can have various substitutions, modifications, and changes without departing from the spirit and scope of the present invention, it is not limited to the embodiments and the accompanying drawings. Judgment should be made including scope and equivalence.

The present invention implements a compression and decompression method with low loss of image information in a simple step, thereby reducing the amount of image data to be processed, thereby processing a smaller amount of data in a video RAM serving as a frame buffer. Improved processing speed

In addition, since the compression rate does not vary depending on the content of the image, the image compression rate of the present invention has a stable processing time of the image data.

Furthermore, the image processing apparatus using the compression and decompression method of the present invention can reduce the amount of image data and improve the processing speed of image information by making encoders and decoders with simple logic circuits because the compression method does not use complex equations.

Claims (13)

Recognizing the current pixel RGB data D _N and the decompressed data D _N-1 ′ based on the previous pixel RGB data D _{N −1} ; Calculating a difference value of the D M _{N N} the D _N-1 'for; If the M _N is positive or zero, the sign bit is recognized as 0, and if the _{N N} is negative, the sign bit is recognized as 1, The M _N is divided into an upper bit and a lower bit, If the upper bit is not 0, the position bit is 1 and the upper bit is recognized as a data bit, Recognizing a location bit as zero and recognizing the lower bit as a data bit if the upper bit is zero; And Image coding method comprising the steps of obtaining the _N D to the sign bit, the bit position, and the mutator bit-compressed data D _N ''. (Where N represents the order of each pixel, where N = 1, 2, 3, ... n, ..., n is a natural number) Recognizing the compressed data D _N ″ for the RGB data of the current pixel and the decompressed data D _N-1 ′ based on the RGB data D _N-1 of the previous pixel; Calculating a sign bit, a position bit, and a data bit from the compressed data D _N ″; If the bit position is 1, the upper bits and the lower bits of the M _N-bit data of the M _N is recognized as 0 If the bit position is 0 and the lower bits of the data bit and the upper bit of the M _N M _N is recognized as 0, Recognizing the M _N as a positive number when the sign bit is 0 and recognizing the M _N as a negative value when 1, restoring the difference value M _N ; And And adding the M _N to the D _N-1 'to obtain the extended data D _N '. Recognizing the decompression data D _N-1 ′ based on the current pixel RGB data D _N and the previous pixel RGB data D _N-1 ; Calculating a difference value of the D M _{N N} the D _N-1 'for; If the absolute value of M _N is greater than the maximum value that H bits (where H represents half of the number of bits occupied by one subpixel RGB data of the one pixel) can be recognized as a position bit 1 The data bit is recognized as a value from the top of the D _N to the H + 1 bit, If the absolute value of M _N is less than or equal to the maximum value that the H bit can have, recognizing a position bit as 0 and recognizing a data bit as a value from a lower part of M _N to a value of H + 1 bits; And Image coding method comprising the steps of obtaining the D position to the _N-bit data and the bit-compressed data D _N ''. Recognizing the compressed data D _N ″ for the RGB data of the current pixel and the decompressed data D _N-1 ′ based on the RGB data D _N-1 of the previous pixel; Calculating position bits and data bits from the compressed data D _N ″; And If the location bit is 1, the upper part of D _N 'is recognized as the value of the data bit from H + 1 bit, and the lower part of D _N ' is recognized as the maximum value that H-2 bit can have from H-1 bit. , When the position bit is 0, the first bit of the data bit is recognized as a sign bit, and the value added by the remaining bit and D _N-1 'is recognized as D _N ', thereby compressing the compressed data D _N ''. Elongated data for D _n 'get Image stretching method comprising the step. The method according to claim 1 or 3, And obtaining compressed data for each pixel RGB data sequentially for all pixels of the screen. The method of claim 5, And recognizing RGB data D ₁ of the first pixel at the top of the screen as extended data D ₁ ′. The method of claim 5, Image compression, characterized in that it comprises a step of recognizing in any RGB data D virtual pixel of having a _zero neighborhood located at the first pixel at the top of the screen and extending the D ₀ of the virtual pixel data D ₀ ' Way. The method according to claim 2 or 4, And obtaining the decompressed data for the compressed data of each pixel RGB data sequentially for all the pixels of the screen. Calculates a difference value M _N between the current pixel RGB data D _N and the extended data D _N-1 ′ based on the previous pixel RGB data D _N-1 , and based on the calculated difference value M _N; An encoder for compressing pixel RGB data D _N of to obtain compressed data D _N ″; A frame buffer for buffering the compressed data D _N ″ calculated by the encoder; And And a decoder for decompressing the compressed data D _N ″ buffered by the frame buffer to obtain the decompressed data D _N ′. The method of claim 9, And the encoder receives the extended data D _N-1 ′ based on the previous pixel RGB data D _N-1 from the decoder. The method of claim 9, And the encoder compresses the RGB data by the image compression method of claim 1 or the image compression method of claim 3. The method of claim 9, The decoder extends the compressed data by the image decompression method of claim 2 or the image decompression method of claim 4. The method of claim 9, And an output unit for outputting an RGB data signal of each pixel of the input image to an encoder, and an output unit for outputting D _N 'extended by the decoder.

Images