KR100782254B1 - Method for detecting word regions using camera - Google Patents

Abstract

The present invention relates to a method of extracting a desired text area when a camera is taken in a document. In particular, the present invention relates to a desired text area when a camera embedded in or mounted on a portable mobile terminal such as a mobile phone is reflected on a document. And a method for extracting a text area using a camera to store the extracted image. As described above, in the method for extracting a text area according to the present invention, a process of extracting a specific text area in a portable mobile terminal equipped with a camera is extracted in real time from a preview screen to display the coordinate area of the specific text.

Camera, Mobile Terminal, Character Extraction, Preview Screen

Description

Character Area Extraction Method Using Camera {METHOD FOR DETECTING WORD REGIONS USING CAMERA}

1 is a flowchart illustrating a text area extraction process using a camera according to the present invention.

2 is a view showing a character region extraction before using a camera according to the present invention;

3 is a view for explaining a method for adjusting the absolute center coordinates of the text area extraction process using a camera according to the present invention

4 is a view for explaining a y coordinate extraction method of the text area extraction process using a camera according to the present invention;

5 is a view for explaining the x-coordinate extraction method of the text area extraction process using a camera according to the present invention

6 is a view showing after extracting a text area using a camera according to the present invention;

The present invention relates to a method of extracting a desired text area when a camera is taken in a document. In particular, the present invention relates to a desired text area when a camera embedded in or mounted on a portable mobile terminal such as a mobile phone is reflected on a document. And a method for extracting a text area using a camera to store the extracted image.

In general, a method of acquiring a document image and recognizing a character included in the document image includes an OCR and a scanner-based character recognition method.

Since the OCR system or the scanner-based character recognition method is a document recognition system, extensive application and hardware resources are required for processing and recognition of document images. Therefore, in a device having limited processor and memory resources, it is difficult to recognize and use an existing OCR system or a scanner-based character recognition method as it is.

In particular, as the camera is introduced into a portable mobile terminal, various technologies have been applied by using an image, and continuous technology development is being made.

In addition, there are a number of objects that the user wants to store through the camera, and the document is one of them.

However, the image data itself has a lot of limitations in terms of its storage because it actually has a larger capacity than the information contained in the image.

In order to solve this problem, it is necessary to store only the minimum area of the document contents that the user needs.

However, although image processing techniques for extracting regions have been developed so far, there are limitations in region extraction algorithms in a limited processor hardware environment, such as a mobile mobile terminal, and in a situation that must be performed in real time, such as a camera preview.

Accordingly, the present invention has been made in order to solve the above problems, it is possible to reduce the amount of calculation without unnecessary calculation through a scenario that accepts an image through a camera to extract a text area, but only a partial area from the center to calculate It provides a text area extraction method using a camera.

A feature of the present invention for achieving the above object is a character region extraction method using a camera for extracting a specific character region on a preview screen in real time, the first region for adjusting the absolute center coordinates for extraction of the character region; step; A second step of extracting a first upper and lower Y coordinate by calculating an upper and lower offset apart from the adjusted center coordinates; And extracting a primary left and right X coordinate in the extracted upper and lower Y coordinate regions.

delete

The first step may include obtaining an x-axis histogram value within a predetermined region from a relative center coordinate of a specific word image reflected by the user in units of lines, and calculating the histogram value of the x-axis direction obtained in units of lines. Projecting the Y-axis to compare the projected values with a threshold value to extract an area with a word, and the difference between the center of the worded area and the relative center value of the specific word image reflected by the user (Δ ) And a value obtained by subtracting the obtained value? From the absolute center Y value.

In addition, the projected values are expressed as 1 when the threshold value is greater than or equal to the threshold value and 0 when the threshold value is less than the threshold value.

The portion marked 1 is a region having a word, and the portion marked 0 is a region without a word.

In addition, the histogram is characterized in that to map the range (range) to the range 0 ~ 31.

The third step may include obtaining a histogram value in the Y direction in a column unit within the upper and lower Y coordinate areas, projecting the histogram value obtained in the column unit on a X-axis by comparing with a threshold value, and Scanning the values projected on the X-axis from left to right in order to obtain a range of left positions that are above the first threshold and last ends of the right positions, and an offset of the nearest blank space left and right from the absolute center X coordinate adjusted within the range. calculating (offset), and subtracting and adding the calculated left and right offset values from the adjusted absolute center X coordinates.

The method may further include determining an area within the range that is not greater than or equal to a threshold, and determining whether the area is a word between words and an area between letters and letters, wherein the determination is divided by a predetermined threshold value. If the value is larger than the predetermined threshold value, the word is determined to be a space representing the word and the word is smaller than the predetermined threshold value, and the value is determined to be a space representing the character and the letter.

In the step of projecting the histogram value obtained in the column unit to the X-axis by comparing with the threshold value, it is expressed as 1 if the value is greater than or equal to the threshold value, and 0 if it is less than the threshold value.

In addition, the secondary Y coordinates are extracted with the extracted primary X and Y coordinates, and the secondary X coordinates are extracted through the extracted Y coordinates.

In addition, the display of the extracted text area is displayed in the form of a box and can be confirmed by the user.

Hereinafter, a method of extracting a text area using the camera of the present invention will be described with reference to the accompanying drawings.

1 is a flow chart illustrating a text area extraction process using a camera according to the present invention.

Referring to FIG. 1, first, a user operates a camera attached to a portable mobile terminal, thereby confirming whether the camera is switched to a preview mode (S100).

As a result of the checking (S100), when the camera is switched to the preview mode, the user switches to the preview mode and simultaneously illuminates a specific area, and at this time, the absolute center coordinates fixed as defaults are located at an arbitrary position of the specific area. The center of gravity is adjusted up, down, left, and right so as to be positioned, and the absolute center coordinates are adjusted within a predetermined area from the absolute center coordinates located at this arbitrary position (S110).

Here, when the camera is illuminated on a specific area, as shown in FIG. 2, the absolute center coordinates of which the specific area is fixed as a default is displayed in a cross shape through the display of the portable mobile terminal.

2 is a view showing before the text area extraction using a camera according to the present invention.

That is, when the user illuminates the camera on an area including the "Protect Optics" part, the center coordinates of the specific area are displayed in a cross shape.

In addition, a method of adjusting the absolute center coordinates for extracting a word of a specific area reflected by the user based on the good center coordinates fixed as the default will be described with reference to FIG. 3.

3 is a view for explaining a method of adjusting the absolute center coordinates of the text area extraction process using a camera according to the present invention.

First, when the user illuminates a specific area using the camera, the specific area is displayed with the absolute center coordinates fixed by default, and the word to be extracted by the user included in this specific area is extracted to check it. You need to know the center coordinates of the words you want to say.

To do this, you must adjust the absolute center coordinates.

The center coordinates of the specific word image reflected by the user are regarded as relative center coordinates, and the x-axis histogram inside a predetermined region is determined from the relative center coordinates of the specific word image in units of lines.

Then, the histogram value in the x-axis direction obtained by the line unit is projected onto the Y-axis, and the projected values are compared with a threshold to extract an area having a word.

In this case, the projected values are compared with a threshold and represented as '1' if the threshold is greater than or equal to '0'.

The portion marked with '1' is a region with a word, and the portion marked with '0' is a region without a word.

In addition, the range of the histogram is generally used in the range of 0 to 255, but if the operation is repeated many times, the amount of calculation has a large influence on the total association, so the histogram range is mapped to 0 to 31. use.

That is, 0 ~ 7 → 0, 8 ~ 15 → 1, 16 ~ 23 → 2, 24 ~ 31 → 3, 32 ~ 39 → 4, 40 ~ 47 → 5, 48 ~ 55 → 6, ...... ....... 248 ~ 255 to 31.

When a region having a word is extracted by comparing the projected values with a threshold value, a difference Δ between a center of the region having a word and a relative center value of a specific word image reflected by the user is obtained.

Then, the absolute center coordinate is adjusted from the absolute value Y to the absolute value Y.

The adjusted absolute center coordinates are represented by Equation 1 as follows.

[Equation 1]

는 단어가 있는 영역의 중심과 상기 사용자가 비춘 특정 단어 영상의 상대적인 중심값의 차(△)를 나타낸다. Where

Denotes the difference Δ between the center of the area where the word is located and the relative center value of the specific word image reflected by the user.

Subsequently, when the absolute center coordinates are adjusted, a character extraction algorithm of a specific region is driven from the adjusted center coordinates (S120).

Here, the text extraction of the specific region may be divided into four stages, or may be divided into more stages.

First, the approximate primary y-area coordinates are obtained, and the primary x-coordinate values are extracted through the primary y-coordinates (S120a to S120b).

Then, the correct second y coordinate is extracted again based on the first x and y coordinates, and the second x coordinates are obtained from the extracted y coordinates (S120c to S120d).

First, y coordinate extraction of the primary region will be described with reference to FIG. 4.

4 is a view for explaining a y coordinate extraction method of the text area extraction process using a camera according to the present invention.

As shown in FIG. 4, the primary y-value of the text area is calculated by calculating an offset that is spaced up and down from the adjusted center coordinates to extract the primary Y-coordinates (S120a).

The histogram values extracted from each line of the x-axis are projected on the y-axis as determined in the absolute center coordinate adjustment, and the determined values are determined based on the measured values. Here, the histogram is mapped to the range of 0 to 31, and the projection value is used. Calculate the offset of the upper y and lower y around this adjusted center coordinate Y.

When the upper first upper Y coordinate and the lower Y coordinate of the extracted character area are expressed by [Equation 2], they are as follows.

[Equation 2]

Subsequently, when the primary upper and lower Y coordinates are extracted, the primary left and right X coordinates are extracted in the extracted primary upper and lower Y coordinate regions (S120b).

5 is a view for explaining the x-coordinate extraction method of the text area extraction process using a camera according to the present invention.

First, the histogram value in the Y direction is calculated in the column unit in the upper and lower Y coordinate regions extracted first.

Then, the histogram value obtained in the column unit is projected on the x-axis by comparing with the threshold.

In this case, the threshold is greater than or equal to '1' and the threshold is less than '0'.

Subsequently, the values projected on the x-axis are scanned sequentially from the left and right ends to obtain a range of the first starting left position and the last ending right position, denoted by '1' which is greater than or equal to a threshold.

Then, within the range, the offset of the nearest blank area left and right from the adjusted absolute center X coordinate is calculated, and the calculated left and right offset value is subtracted from the adjusted absolute center X coordinate and added to the primary left. Extract the right X coordinate.

Here, an area less than the threshold value is extracted within the above range, and the extracted area is distinguished between an area between words and words or between letters and letters by another preset threshold value.

In this case, when the division is larger than the threshold, the division is determined to be a space between the words and the word.

When the extracted first left and right X coordinates are expressed by Equation 3 as follows.

[Equation 3]

       Subsequently, the secondary Y coordinates are accurately found again with the first and second extracted X and Y coordinates, and the second X coordinate region is extracted based on the extracted second Y coordinate region (S120c to S120d).

      Here, in the step of extracting the secondary region, the connection is not found in the primary, and the segmentation error is filtered out through a predetermined threshold value to adjust the more accurate final coordinate region (S130). If present, the existing area is displayed as a square. If there is no area, a cross is displayed.

That is, the extracted text area may be displayed in a box (square) form as shown in FIG. 6 so that the user may check the extracted text.

Therefore, the initial preview screen shows a cross-section at a predetermined position through FIG. 2, and when a text image enters the center of the display area, an algorithm is driven to extract a coordinate as in the method described in FIGS. 3 to 5. Display the coordinate area in the form of a box on the screen.

As described above, the present invention receives an image through a camera and extracts a text area, but only a part of the area can be calculated from the center so that unnecessary calculation is not performed.

In addition, when the histogram analysis for character area extraction is used by mapping from 0 to 31 range, the calculation amount is reduced to 8 / 1.Also, the algorithm can be performed in real time on the preview screen showing 15 frames per second and display on the screen. Do.

In addition, even though the histogram is lowered, the center coordinate adjustment and 1, 2 X, Y coordinates have the effect of extracting the text area of the part exactly desired by the user.

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 defined not only by the scope of the following claims, but also by the equivalents of the claims.

Claims (14)

Character area extraction method using a camera for extracting a specific character area from the preview screen in real time, Adjusting an absolute center coordinate for extracting a text area; A second step of extracting a first upper and lower Y coordinate by calculating an upper and lower offset apart from the adjusted center coordinates; And And a third step of extracting first left and right X coordinates in the extracted up and down Y coordinate regions. delete The method of claim 1, wherein the first step, Obtaining a x-axis histogram value within a predetermined area from a relative center coordinate of the specific word image reflected by the user in units of lines; Projecting the histogram value in the x-axis direction obtained on a line-by-line basis to a Y-axis, and extracting a region having a word by comparing the projected values with a threshold value; Obtaining a difference Δ between a center of the area where the word is located and a relative center value of the specific word image reflected by the user; And obtaining a value obtained by subtracting the obtained value Δ from an absolute center Y value. The method of claim 3, wherein And comparing the projected values with a threshold value as 1 when the threshold value is greater than or equal to 1, and displaying the projected values as 0 when the threshold value is less than the threshold value. The method of claim 4, wherein The portion denoted by 1 is a region having a word, and the portion denoted by 0 is a region without a word. The method of claim 3, wherein The histogram is a character region extraction method using a camera, characterized in that for mapping the range (range) to the range 0 ~ 31. The method of claim 1 wherein the third step is Obtaining a histogram value in the Y direction in a column unit within an upper Y coordinate region; Projecting the histogram value obtained in units of columns to an X-axis by comparing with a threshold value; Scanning the values projected on the X axis from left and right ends in order to obtain a left position that is greater than or equal to a threshold that is initially started and a right position range that ends last; Calculating an offset of the nearest blank area left and right in the adjusted absolute center X coordinate within the range; And subtracting and adding the calculated left and right offset values from the adjusted absolute center X coordinates. The method of claim 7, wherein And searching for an area within the range that is not greater than or equal to a threshold, and determining whether the area is a word between words or an area between letters and letters. The method of claim 8, The determination method is a character area extraction method using a camera, characterized in that divided by a predetermined threshold value. The method of claim 9, The method of extracting a character region using a camera according to claim 1, wherein if the value is greater than the predetermined threshold value, the space between the words is determined. The method of claim 9, If the value is less than the predetermined threshold value, the character space extraction method using a camera, characterized in that further comprising the step of re-determining the value to 1 as a space indicating between the character. The method of claim 7, wherein In the step of projecting the histogram value obtained in the column unit on the X-axis in comparison with a threshold value, if the threshold value is greater than or equal to 1, the character region extraction method using a camera characterized in that it is displayed as 0. The method of claim 1, Extracting a secondary Y coordinate with the extracted primary X, Y coordinates, and extracts the secondary X coordinates through the extracted Y coordinates. The method of claim 1, And extracting the text area, and displaying the extracted text area in the form of a box.

Images