KR20080045842A - Apparatus and method for image compression and restoring - Google Patents

Abstract

An apparatus and a method for image compression and restoring are provided to have high-speed compression and restoring and to increase compression efficiency by applying the improved LZSS(Lempel Ziv Storer Szymansky) manner, using the binary-tree retrieval manner and predicting the compression technology. An interface module(310) is a module for communicating with another module and includes an RF(Radio Frequency) processing member and a baseband processing member and the like. A control member(320) controls the general operations of a portable terminal and controls a compression managing member(340) so that the image compression and decompression processes of an image are performed. The compression managing member uses an interframe prediction manner, an improved LZSS manner and a selective compression manner when performing the compression and decompression of plural images. The improved LZSS manner uses the binary tree. A storing member(330) stores a program for controlling the general operations of the apparatus and the temporary data such as the image frames. The control member also performs the functions of the compression and decompression member.

Description

 Apparatus and method for compressing and restoring images {APPARATUS AND METHOD FOR IMAGE COMPRESSION AND RESTORING}

1 is a view showing a typical LZ77 compression scheme,

2 is a diagram illustrating an image compression method according to an embodiment of the present invention;

3 is a view showing a portable terminal according to an embodiment of the present invention;

4 is a view showing an image compression process according to an embodiment of the present invention, and,

5 is a diagram illustrating an image restoration process according to an embodiment of the present invention.

The present invention relates to image compression and decompression, and more particularly, to an apparatus and a method for increasing compression speed and increasing compression efficiency in image compression by finding similarities in a plurality of images used in an application of a portable terminal.

Applications of various portable terminals, including graphical user interfaces, contain a large number of graphical images. As the function of the portable terminal is improved, the function of the application program is also improved, and accordingly, the number of graphic images required is also increasing.

As the number of graphic images increases, there is a need for a compression technique for graphic images. High compression efficiency and fast compression and decompression speeds are fundamental to compression. The former can obtain small sized images, while the latter can achieve fast speeds. However, lossy compression techniques with high compression efficiency may not achieve fast speeds due to slow compression and decompression. Also, among non-lossy compression algorithms, entropy coding (eg, Huffman coding) takes advantage of probability distribution, but it is difficult to obtain high compression efficiency because it ignores positional relations (in terms of time and space). There is this.

The LZ77 compression method, which is based on the LZW method, has been widely used since the patent expired, but since it is based on the full search method, it takes a long time to compress all the images in the portable terminal.

1 illustrates a general LZ77 compression method.

Referring to FIG. 1, the LZ77 scheme compresses only the overlapping portion within one frame as shown in FIG. 1A. Also, as shown in FIG. 1B, the position of the matched string in the search buffer uses an absolute position value.

As described above, the method of compressing only the overlapping portion in one frame does not consider overlapping between frames, which causes inefficiency and low compression efficiency.

It is an object of the present invention to provide an apparatus and method for image compression and reconstruction.

Another object of the present invention is to use an image compression and reconstruction method that improves the LZSS method using a method of compressing a first frame using a binary tree search method and a method of compressing a next frame using an inter-frame prediction technique. An object of the present invention is to provide an apparatus and method for increasing compression and restoration speed and increasing compression efficiency.

According to a first aspect of the present invention for achieving the above object, in the method for compressing an image, after loading the first image frame, a portion of a predetermined size or more matched within the first image frame is displayed using a binary tree. A first compression process of compressing the first image frame, a second compression process of compressing the first image frame, and then loading the next image frame to display and compress a matching portion with the first image frame using a binary tree; And after the second compression process, a third compression process of displaying and compressing a portion having a predetermined size or more matched in the next image frame by using a binary tree.

According to a second aspect of the present invention for achieving the above object, in a method for reconstructing a compressed image, after loading the first compressed image frame, copying and reconstructing the matching portion in the first image frame A first restoring process, a second restoring process of restoring the first image frame, and then restoring by copying a matching portion with the first image frame by loading a compressed next image frame, and after the second restoring process, And a third restoration process of copying and restoring a matched portion in the next image frame.

 According to a third aspect of the present invention for achieving the above object, in the apparatus for compressing and restoring an image, during image compression, a portion of a predetermined size or more matched within the first image frame is displayed and compressed using a binary tree. Load the next image frame to display and compress a portion matching the first image frame by using a binary tree, and display and compress a portion of a predetermined size or more matched within the next image frame by using a binary tree, and compress the image. In restoring, copying and restoring a matched portion in the first compressed image frame, loading the next compressed image frame, copying and restoring a portion matching the first compressed image frame, and restoring in the next compressed image frame Compression management unit for copying and restoring the matching part in the, It characterized in that it comprises compression group management unit and the input-output storage for storing a compressed image and an uncompressed image, necessary data when compressed or restored portion.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, when it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

The present invention will now be described with respect to apparatus and methods for image compression and reconstruction.

The improved LZSS compression scheme in the present invention uses a binary tree search scheme using a ring buffer. First, the LZSS scheme of the present invention based on binary tree search will be described.

2 is a diagram illustrating an image compression method according to an exemplary embodiment of the present invention.

Referring to FIG. 2, the improved LZSS method of the present invention uses a ring buffer consisting of a search buffer, a lookahead buffer, and a tail of a buffer as shown in FIG. 2B. do.

First, some strings are read from the image and stored in the look ahead buffer. The longest string is then retrieved from the search buffer and the matching string position and length (in the ring buffer) of the longest string and the string just loaded from the image are output. Image data is loaded in the search buffer.

If the longest match is two bytes long, only one character is output without encoding, and then the above process is repeated for the next data.

If the length of the ring buffer is 2048 bytes, the position of the matching string may be represented by 11 bits. When the length of the matched string is represented by 5 bits and output using the display format of <position, length>, 2 bytes are required.

When only one character without the display format or encoding of <position, length> is output, one extra bit is output and used as a separator. The display format of <position, length> is stored in a binary tree structure. The binary tree has three arrays (dad [], lson [] and rson []) and does not perform pre-stuffing to retrieve the buffer. Since the search operation is to search a binary tree, the search time can be significantly reduced.

All <position, length> display formats are relative and not absolute, as shown in FIG. 2B. That is, the next value is displayed based on the previous value. Therefore, it can be restored immediately without creating a ring buffer like compression.

When the improved LZSS scheme of the present invention is used to compress the first image among a plurality of images, the output code package for outputting the compressed image is composed of one leading flag byte and eight code units. The output code package may be embedded in a header of an image frame. Each bit in the flag byte that points to the next corresponding unit represents a <position, length> (denoted 0) or an unencoded character (denoted 1).

The position ranges from 0 to 2047 (11 bits) and the length ranges from 3 to 34 (5 bits). Thus, the <position, length> representation format is " 11 bits + 5 bits ", that is, 2 bytes in length, and the character without encoding is 1 byte in length.

In the present invention, when compressing a plurality of image frames, the inter-frame prediction and the improved LZSS scheme are merged and used as shown in FIG. 2A. This will be described below.

The above scheme considers two types of matching. One is the match of the longest string between the current frame and the previous frame and the other is the match of the longest string in the current frame. In order to avoid time consuming during full search and to take advantage of gradual image changes, interframe prediction is performed by comparing strings of the same position between adjacent frames. Therefore, the output of interframe matching has only the longest matching length and the matching position is omitted. The matching length may be selectively indicated by allocating 1 byte or 2 bytes.

The improved LZSS scheme of the present invention is also used in the matching process in the current image. The output of the match in the current image is <position, length>. Outputs the longer of a match between the images and a match in the image.

Of course, when outputting interframe matching or intraframe matching, an extra bit is output and used as a separator.

The interframe prediction technique uses a selective compression method. The inter-frame prediction method has an advantage of increasing compression efficiency when there are many similarities between frames.

All successive image frames are compressed using the merging technique above. The output code package is composed of one reading flag byte and eight code units as described above. The output code package may be embedded in a header of a compressed image frame. Each bit in the flag byte indicating the next corresponding unit represents a &quot; position, length &quot; (denoted by 0) or a character without encoding (denoted by 1).

However, the matching unit (<position, length> display format) is configured differently. The matching unit is composed of one or two bytes, and the first bit indicates whether a match between frames (indicated by 1) or in-frame matching (indicated by zero).

In the case of in-frame matching, one flag bit is shortened to indicate in-frame matching. The position ranges from 0 to 2047 (11 bits) and the length ranges from 3 to 18 (4 bits). This indicates that the longest matching length of the next frame is half of the longest matching length of the first frame. As a result, the <position, length> display format is "1 flag bit (denoted by 0) + 11 bits + 4 bits", that is, 2 bytes in length.

In the case of the interframe matching, the first bit is used to indicate interframe matching (indicated by 1). The second bit indicates whether the matching length is represented by 6 bits (3-63) (denoted by 0) or by 14 bits (64 to 16383) (denoted by 1).

As a result, the <Length> display format is "1 flag bit (marked as 1) + 1 range bit (marked as 0) + 6 bits", or 1 byte long or "1 flag bit (marked as 1) + 1 range bit ( 1 bit) + 14 bits ", 2 bytes long.

3 is a block diagram of a portable terminal according to an embodiment of the present invention.

Referring to FIG. 6, the interface module 310 is a module for communicating with other nodes, and includes an RF processor and a baseband processor. The RF processor converts a signal received through an antenna into a baseband signal and provides the baseband signal to the baseband processor, and transmits a baseband signal from the baseband processor on an actual air. Change to and transmit through the antenna.

The controller 320 controls the overall operation of the portable terminal. For example, processing and control for voice call and data communication are performed, and in addition to the usual functions, the compression management unit 340 is controlled according to the present invention to perform image compression and reconstruction.

The compression management unit 340 uses an improved LZSS method that does not regenerate a ring buffer when compressing and restoring an image according to instructions and provision information of the control unit 320, and predicts interframes when compressing and restoring a plurality of images. Scheme and the improved LZSS scheme and selective compression scheme.

The storage unit 330 stores a program for controlling the overall operation of the apparatus and temporary data generated during program execution. In particular, it stores image frames according to the invention.

In the above block configuration, the control unit 320 may perform a function of the compression management unit 340. In the present invention, it is shown to configure them separately to explain each function separately. Accordingly, when the product is actually implemented, all of the functions of the compression manager 340 may be configured to be processed by the controller 320, and only some of the functions may be configured to be processed by the controller 320.

4 is a diagram illustrating an image compression process according to an exemplary embodiment of the present invention.

Referring to FIG. 4, first, an image is loaded (step 410). Then, it is checked whether the image is the first frame (step 420).

If the first frame, the length of the look ahead buffer is set to 34 (step 430), and compression is performed using the improved LZSS scheme of the present invention (step 460). An improved LZSS scheme is to use binary trees.

If it is not the first frame but the next frame, the length of the look ahead buffer is set to 18 (step 440), using the interframe prediction technique of the present invention (step 450), and the improved LZSS scheme of the present invention. Compression is performed (step 460).

Then, the algorithm according to the present invention is terminated.

5 is a diagram illustrating an image restoration process according to an embodiment of the present invention.

Referring to FIG. 5, the first compressed image is reconstructed using the improved LZSS method (step 510).

Thereafter, the second image is reconstructed using an interframe prediction method of copying the same longest string of the character without encoding and the previous image to the same position of the current image according to the flag bit (step 520). Even in this case, the improved LZSS scheme is used. From the third image, the above process is repeated and restored.

Then, the algorithm according to the present invention is terminated.

Although the technique of the present invention is not shown in the present invention, according to the experiment, the compression speed is about 20 times faster than the conventional technology, and the compression efficiency is higher by about 15% (LZ77: 41%, the method of the present invention: 26%). ..

Meanwhile, in the detailed description of the present invention, specific embodiments have been described, but various modifications are possible without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the scope of the following claims, but also by the equivalents of the claims.

The compression method of the present invention has a fast speed because it restores a compressed image without generating a ring buffer during restoration. In addition, since the compression method using the inter-frame prediction is used, the compression efficiency is high.

Claims (25)

In the method for image compression, A first compression process of loading a first image frame and displaying and compressing a portion of a predetermined size or more matched within the first image frame using a binary tree; A second compression process of compressing the first image frame and then loading the next image frame to display a portion matching the first image frame using a binary tree and to compress the first image frame; And after the second compression process, a third compression process of displaying and compressing a portion having a predetermined size or more matched in the next image frame by using a binary tree. The method of claim 1, The second compression process may include compressing by displaying a length of a portion matched with the n th image frame when the n + 1 th image frame is compressed. The method of claim 1, The third compression process includes, when the n + 1 th image frame is compressed, displaying and matching a matching portion within the n + 1 th image frame. The method of claim 1, And wherein the indication of the matched portion in the first compression process is "(matching position, matching length)". The method of claim 4, wherein Wherein "(matching position, matching length)" includes information indicating a matching position and information indicating a matching length. The method of claim 1, And wherein the indication of the matched portion in the second compression process is "(matching length)". The method of claim 6, And "(matching length)" includes information indicating matching between image frames, information indicating a range of matching lengths, and information indicating a matching length. The method of claim 1, And wherein the indication of the matched portion in the third compression process is "(matching position, matching length)". The method of claim 8, Wherein "(matching position, matching length)" includes information indicating a match in an image frame, information indicating a matching position, and information indicating a matching length. The method of claim 1, The matching portion having the predetermined size or less is not compressed in the first compression process and the third compression process. In a method for restoring a compressed image, A first restoration process of copying and restoring a matched portion within the first image frame after loading the first compressed image frame; Restoring the first image frame, loading a second compressed image frame, and copying and restoring a matching portion with the first image frame; And a third restoration process of copying and restoring a matched portion in the next image frame after the second restoration process. The method of claim 11, The second restoring process includes copying and restoring a matching portion with the n th image frame when restoring the n + 1 th image frame. The method of claim 11, The third restoring process may include copying and restoring a matching portion within the n + 1 th image frame when restoring the n + 1 th image frame. The method of claim 11, The matching portion in the first restoration process is characterized in that "(matching position, matching length)". The method of claim 14, And " (matching position, matching length) " is information indicating a matching position and information indicating a matching length. The method of claim 11, The matching portion in the second restoration process is characterized in that "(matching length)". The method of claim 16, And "(matching length)" includes information indicating matching between image frames, information indicating a range of matching lengths, and information indicating a matching length. The method of claim 11, The matching portion in the third restoration process is characterized in that the "(matching position, matching length)". The method of claim 18, Wherein "(matching position, matching length)" includes information indicating a match in an image frame, information indicating a matching position, and information indicating a matching length. In the apparatus for compressing and restoring an image, When compressing an image, a portion of a predetermined size or more matched within the first image frame is displayed and compressed using a binary tree, and the next image frame is loaded and displayed and compressed by using a binary tree. And compressing by displaying a portion of the matching size or more within the next image frame using a binary tree, When restoring an image, copying and restoring a matched portion in the first compressed image frame, loading the next compressed image frame, copying and restoring a portion matching the first compressed image frame, and restoring the next compressed image frame A compression management unit for copying and restoring a matching portion in the system; The compression management unit comprises a storage unit for inputting and outputting data required for compression or decompression, and for storing the compressed image and the uncompressed image. The method of claim 20, The compression management unit compresses by displaying a length of a portion matched with the n-th image frame when compressing the n + 1 th image frame. The method of claim 20, When the compression management unit compresses an n + 1 th image frame, the compression management unit displays and compresses a matching portion within the n + 1 th image frame. The method of claim 20, And when the n + 1 th image frame is to be restored, the compression management unit copies and restores a matching portion with the n th image frame. The method of claim 20, When the compression management unit restores the n + 1 th image frame, the compression management unit copies and restores a matching portion within the n + 1 th image frame. The method of claim 20, And the compression management unit does not compress matching portions of a predetermined size or less.

Images