KR20050106810A - Method and apparatus for processing of binary image - Google Patents

Abstract

A binary image processing method and apparatus are disclosed. The binary image processing method according to the present invention comprises the steps of dividing a binary image applied from an image source into a predetermined region, and determining whether the image constituting each divided region is a symbol image or a non-symbol image, and comprising a symbol image. Searching for a symbol existing in the symbol image area for areas determined to be traced, and tracking the outline of the retrieved symbol, and extracting the completed symbol from the binary image when the trace of the outline is completed. This can reduce the occurrence of singletons in binary image compression.

Description

Binary image processing method and apparatus {Method and apparatus for processing of binary image}

The present invention relates to a binary image processing method and apparatus, and more particularly, to a binary image processing method and apparatus for extracting symbols from a binary image for efficient encoding when a binary image is to be compressed by a symbol matching coding scheme. .

As a lossless compression method for binary images, Modified Huffman (MH), Modified READ (MR), Modified Modified READ (MMR), and Joint Bi-level Image experts Group (JBIG) are applied. Among them, MR and MMR methods are encoding algorithms applied to G3 and G4 fax machines. JBIG is an arithmetic encoding algorithm based on a context. Recently, JBIG2, which is capable of lossy and lossless encoding, has been standardized as a coding method for binary images, and JBIG2 employs an encoding method based on symbol matching.

Referring to the encoding scheme based on symbol matching, first, a symbol is extracted from an input binary image, and a symbol similar to the extracted symbol is searched for in a dictionary or a library. At this time, the image extracted as a symbol means an image such as text.

If it is determined that a symbol similar to the extracted symbol exists in the dictionary or library, the information is encoded using index information of a symbol stored in the dictionary. On the other hand, if a symbol similar to a previously extracted symbol does not exist, the extracted symbol is encoded by newly registering.

The encoding scheme based on the symbol matching is efficient in encoding image data divided into symbols, such as text and symbols, but is inefficient because it lowers the compression ratio in encoding image data such as line-art or halftone images. . In general, a document composed of binary images contains images separated by symbols such as text and symbols, and images separated by nonsymbols such as line art and halftone images. Therefore, when encoding a document in which an image divided by a symbol and an image divided by a non-symbol are mixed by using an encoding method based on symbol matching, the compression ratio is lowered as a whole, which is inefficient.

In order to solve this problem, binary images are divided into partial images having predetermined characteristics and then compressed by using methods such as RLSA (Run Length smearing Algorithm), RXYC (Recursive X-Y Cut), and Docstrum. However, since these methods require a large amount of computation or a large amount of memory, they are mainly applied to the fields of character recognition or document structure interpretation.

On the other hand, when dividing a binary image in the above manner, symbols exist at the boundary of the divided image. Due to these symbols, a singleton is generated during the compression process, and the compression rate of the binary image is reduced. There are disadvantages.

SUMMARY OF THE INVENTION The present invention has been made in an effort to provide a method and apparatus for extracting a symbol from a binary image which efficiently extracts symbols used for symbol matching from the binary image for efficient encoding of the binary image.

In order to solve the above technical problem, the binary image processing method according to the present invention divides a binary image applied from an image source into a predetermined region, and determines whether the image constituting each of the divided regions is a symbol image or a non-symbol image. Determining cognition; Searching for a symbol existing in the symbol image area for regions determined to be composed of the symbol image; And tracing the outline of the retrieved symbol, and extracting the traced symbol from the binary image when the outline tracking is completed.

Preferably, the method further includes determining whether the contour tracking object symbol is included in the non-symbol image area, and wherein the extracting step is determined to include the contour tracking object symbol in the non-symbol image area. If so, the non-symbol image area is traced.

The method may further include searching for whether a similar symbol corresponding to the extracted symbol exists in a previously created dictionary; And if the similar symbol corresponding to the extracted symbol does not exist in the dictionary, adding the extracted symbol to the dictionary.

On the other hand, the binary image processing apparatus according to the present invention for solving the above technical problem, divides the binary image applied from the image source into a predetermined region, and from the divided respective regions to the image constituting the respective region. An information extracting unit for extracting at least one information about the unit; An image determining unit determining whether an image constituting each block is a symbol image or a non-symbol image based on the extracted at least one information; And searching for a symbol existing in a region of the binary image, the symbol being included in the symbol image, tracking the outline of the retrieved symbol, and completing the tracking from the binary image when the contour tracking is completed. It has a symbol extraction unit for extracting.

Preferably, the symbol extractor determines whether the contour tracking target symbol is included in the non-symbol image area, and tracks the non-symbol image area if it is determined that the symbol is included in the non-symbol image area.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings.

1 is a schematic block diagram of a binary image processing apparatus according to a preferred embodiment of the present invention.

Referring to FIG. 1, the binary image processing apparatus 100 includes a scanner 110, a preprocessing unit 120, an information extracting unit 130, an image determining unit 140, a symbol extracting unit 150, and a first compression unit. And a second compression unit 170.

The binary image processing apparatus 100 according to the present invention efficiently extracts the symbols suitable for the symbol matching coding scheme from the binary image in order to efficiently encode the binary image read by the image reading apparatus such as the scanner 110.

The scanner 110 scans a document to be scanned and converts it into a digital signal and outputs the converted digital signal. The binary image converted into a digital signal by the scanner 110 is provided to the preprocessor 120.

The preprocessor 120 performs preprocessing such as noise filtering and tilt correction on the binary image received from the scanner 110.

The information extracting unit 130 divides the preprocessed binary image into predetermined regions, and extracts at least one information for determining characteristics of the image constituting each region from the divided regions. The information extracted by the information extracting unit 130 may include at least one of the number of connected components in each region and color change rate information between pixels in each region. At least one information extracted by the information extracting unit 130 is provided to the image determining unit 140.

The image determining unit 140 determines whether an image constituting each region is a symbol image or a non-symbol image based on the connection element number information and the color change rate information extracted by the information extracting unit 130. Here, the symbol image means an image divided into texts such as text, symbols, and numbers, and the non-symbol image means an image such as a halftone image. In the present invention, an image not determined to be a symbol image is determined to be a non-symbol image. Patent Application No. P2004-0027983, previously filed by the same applicant, for a method of determining whether an image constituting each area divided into predetermined areas by the information extraction unit 130 is a symbol image or a non-symbol image. Since it is disclosed in the detailed description thereof will be omitted.

The symbol extractor 150 performs a symbol search operation included in the binary image, but performs the symbol search operation only on a region determined as a symbol image region among the divided regions. The symbol extractor 150 extracts the symbols existing in the region determined as the symbol image region. When the symbol is searched in the symbol image area, the symbol extraction unit 150 traces the outline of the found symbol, and when the trace of the outline is completed, extracts the completed symbol from the binary image. At this time, if it is determined that the symbol under contour tracking is included in the non-symbol image area, that is, if it is determined that the symbol spans between the symbol image area and the non-symbol image area, the symbol extracting unit 150 is under contour tracking. Traces up to the non-symbol image area containing the symbol. Therefore, according to the present invention, not only the symbol existing in the symbol image region but also the symbol existing between the symbol image region and the non-symbol image region are extracted. However, symbols existing in the non-symbol image area are not extracted. The symbols extracted by the symbol extraction unit 150 are provided to the first compression unit 160.

The first compression unit 160 is a module that compresses the symbols received from the symbol extraction unit 150 and performs a compression process using a compression algorithm based on symbol matching. Here, the compression algorithm based on symbol matching may be Joint Bi-level Image experts Group-2 (JBIG2).

The second compression unit 170 compresses the remaining image in which the symbols are excluded from the binary image, that is, the image existing in the region determined as the non-symbol image region. Compression algorithms applicable to the second compression unit 170 include MR (Modified READ), MMR (Modified Modified READ), halftone coding scheme, JBIG1, and the like.

FIG. 2 is a flowchart illustrating an image processing method in the binary image processing apparatus shown in FIG. 1.

1 and 2, when a binary image is received from the scanner 110 (S210), the preprocessor 120 performs preprocessing such as noise filtering and tilt correction on the received binary image (S215). .

The information extracting unit 130 divides the preprocessed binary image into predetermined regions (S220), and extracts at least one information for determining an image constituting each region from the divided regions. The information extracting unit 130 extracts the number of connection elements in the region and / or color change rate information between pixels existing in the region from each region.

The image determining unit 140 determines whether an image constituting each divided region is a symbol image or a non-symbol image based on the at least one information extracted by the information extracting unit 130 (S225).

When the image determining unit 140 determines the type of the image constituting each area, the symbol extraction unit 150 determines an area determined as being composed of a symbol image among the divided areas, that is, an area determined as a symbol image area. The symbol existing within is searched for (S230). As shown in FIG. 3, the symbol extractor 150 searches for a symbol existing in the symbol image region while moving the region determined as the symbol image region in the raster scan direction. 3 is a diagram illustrating a binary image labeled as a result determined by the image determining unit 140. In FIG. 3, areas labeled with 'T' are areas that are determined to be composed of text data, meaning symbol image areas, and areas labeled with 'H' and 'I' respectively represent halftone images and the like. An area judged to be composed of an intermediate image, which means a non-symbol image area.

If the symbol is found in step S235, the symbol extraction unit 150 tracks the outline of the found symbol in a set direction (eg, clockwise direction) (S235). If the symbol being traced is a symbol that spans between the symbol image region and the non-symbol image region, the symbol extractor 150 performs tracking across the non-symbol image region where the searched symbol exists (S240, S245). That is, as shown in FIG. 4, when the retrieved symbol "S" exists between the symbol image region and the non-symbol image region, the symbol extractor 150 may include the non-symbol image region. Perform tracing until.

When contour tracking for the retrieved symbol is completed (S250), the symbol extractor 150 extracts the completed symbol from the binary image (S255). FIG. 5 shows an image in which an “S” symbol is extracted from the binary image shown in FIG. 4. As shown in Fig. 5, the symbol whose contour tracking has been completed is removed from the binary image. The symbols finally extracted by the symbol extraction unit 150 are provided to the first compression unit 160.

The first compression unit 160 matches the extracted symbols with pre-registered similar symbols and compresses the extracted symbols (S260). Referring to the symbol matching process, it is searched whether a similar symbol corresponding to a symbol extracted by the symbol extraction unit 150 exists in a previously created dictionary. If there is a similar symbol corresponding to the previously extracted symbol, the extracted symbol is compressed by using index information of the previously registered similar symbol. In contrast, if it is determined that there is no similar symbol corresponding to the previously extracted symbol, the extracted symbol is determined as a new symbol and registered in advance. In this way, the compression process is also performed on the positional information on the space of the extracted symbol during the extracted symbol compression process.

6 is a diagram illustrating an example of a non-symbol image compressed by the second compression unit 170. As illustrated in FIG. 6, an image from which a symbol image extracted by the symbol extraction unit 150 is removed from a binary image, that is, a non-symbol image, is compressed by the second compression unit 170.

While the above has been shown and described with respect to preferred embodiments of the present invention, the present invention is not limited to the specific embodiments described above, it is usually in the technical field to which the invention belongs without departing from the spirit of the invention claimed in the claims. Various modifications can be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or the prospect of the present invention.

As described so far, according to the present invention, a binary image is divided into a region composed of a symbol image and a region composed of a non-symbol image, and is present between symbols included in the symbol image region and between a symbol image region and a non-symbol image region. By extracting and compressing only the symbols to be able to accurately match the symbols, it is possible to reduce the occurrence of a singleton.

1 is a schematic block diagram of a binary image processing apparatus according to a preferred embodiment of the present invention;

2 is a flowchart illustrating an image processing method in the binary image processing apparatus shown in FIG. 1;

3 shows a binary image labeled as a result judged by the image determining unit;

4 is a view illustrating a symbol tracking process of a symbol extraction unit;

FIG. 5 is a diagram illustrating an image from which a symbol whose tracking has been completed in FIG. 4 is extracted;

6 is a diagram showing an example of a non-symbol image compressed by the second compression unit.

Explanation of symbols on the main parts of the drawings

100 binary image processing device 110 scanner

120: preprocessing unit 130: information extraction unit

140: image determination unit 150: symbol extraction unit

160: first compression unit 170: second compression unit

Claims (5)

Dividing a binary image applied from an image source into a predetermined region, and determining whether the divided image is a symbol image or a non-symbol image; Searching for a symbol existing in the symbol image area for regions determined to be composed of the symbol image; And Tracing the contour of the retrieved symbol and extracting the trace-completed symbol from the binary image when the contour tracing is completed. The method of claim 1, Determining whether the contour tracking target symbol is included in the non-symbol image area; And the extracting step tracks the non-symbol image region if it is determined that the contour tracking target symbol is included in the non-symbol image region. The method of claim 2, Searching whether there is a similar symbol corresponding to the extracted symbol in a previously created dictionary; And And adding the extracted symbol to the dictionary when the similar symbol corresponding to the extracted symbol does not exist in the dictionary. An information extraction unit for dividing a binary image applied from an image source into a predetermined area and extracting at least one information about an image constituting each area from the divided respective areas; An image determining unit determining whether an image constituting each block is a symbol image or a non-symbol image based on the extracted at least one information; And Search for a symbol existing in a region of the binary image determined to be composed of the symbol image, trace the outline of the retrieved symbol, and if the contour tracking is completed, extract the completed symbol from the binary image. And a symbol extracting unit to extract the binary image processing apparatus. The method of claim 4, wherein The symbol extracting unit determines whether the contour tracking target symbol is included in the non-symbol image area, and if it is determined that the symbol is included in the non-symbol image area, tracks the binary image to the non-symbol image area. Processing unit.