KR101128185B1 - Character recognition apparatus and method of camera phone - Google Patents

Abstract

BACKGROUND OF THE INVENTION Field of the Invention The present invention relates to a character recognition apparatus and method for automatically recognizing characters through a camera without an automatic focusing function.

Character recognition apparatus of a camera phone according to an embodiment of the present invention, the image capturing unit for image capture; A text area recognition unit for activating an automatic text recognition function after checking a degree of coincidence of a specific area of a plurality of continuous frames inputted from the image capturing unit and a focus state of the plurality of continuous frames; An image capture unit for capturing an image when a character recognition function is activated by the character area recognition unit; Characterized in that it comprises a character recognition unit for recognizing the character of the image captured by the image capture unit.

Camera, phone, character recognition

Description

Character recognition apparatus and method of camera phone

1 is a view showing a character recognition method in a conventional camera phone.

2 is a block diagram showing a character recognition device in a camera phone according to an embodiment of the present invention.

3 is a detailed configuration diagram for area fixing notification of the text area recognition unit according to the present invention.

4 shows a Laplacian mask according to the invention.

5 and 6 is a view comparing the frequency domain of the images taken with the camera phone according to the present invention.

7 is a flow chart showing a character recognition method in the camera phone according to the present invention.

<Description of the symbols for the main parts of the drawings>

10 ... image capturing unit 20 ... character area recognition unit

30 ... Image Capture 40 ... Character Recognition

BACKGROUND OF THE INVENTION Field of the Invention The present invention relates to a character recognition apparatus and method for automatically recognizing characters through a camera without an automatic focusing function.

In recent years, with the development of information and technology, various types of terminals have been introduced, which are beyond the simple function of making a phone call and have other additional functions. In addition, portable terminals equipped with cameras of hundreds of megapixels or more have been released, and are being developed into multimedia terminals such as video calls and broadcast viewing.

The portable terminal has a built-in camera or externally installed, and a terminal without an automatic focusing function has been released. In order to recognize texts on these terminals, it is necessary to capture clear and clear images, but it is impossible to recognize accurate texts.

However, if the user manually focuses with his eyes, the accuracy is lowered, and it is inconvenient for the user to attempt shooting by pressing a key button.

1 is a flowchart illustrating a character recognition method in a conventional portable terminal.

Referring to FIG. 1, when an image is input from a camera (S1), a frequency component is obtained for the input image (S2).

At this time, it is checked whether the obtained frequency value is greater than or equal to the threshold value (S3), and when the result of the check result has a frequency value greater than or equal to the threshold value, the corresponding area is considered to be a character having a high frequency component to obtain focus. The focused image is captured and automatically recognized as a text (S5, S6).

However, in such a case, there may be a problem that the user simply moves to the character recognition stage by simply having a high frequency component where it is not desired.

It is a first object of the present invention to provide a character recognition apparatus and method for automatically recognizing a character when a character is in close proximity to a camera.

It is a second object of the present invention to provide a character recognition apparatus and a method for acquiring and calculating only a specific area of a center portion of every frame.

A third object of the present invention is to provide a character recognition apparatus and method for improving a character recognition rate since an image having a well-focused image is received.

A fourth object of the present invention is to automatically recognize a character by automatically recognizing a character through a focus notification and a screen lock function when a user takes the camera phone to a character without a key input in a mechanically non-auto focusing phone. An object and apparatus for character recognition are provided.

Character recognition apparatus in the camera phone according to the present invention for achieving the above object,

In the character recognition device of the camera phone,

An image capturing unit for capturing an image;

A text area recognition unit for activating an automatic text recognition function after checking a degree of coincidence of a specific area of a plurality of consecutive frames inputted from the image capturing unit and a focus state of the continuous predetermined frames;

An image capture unit for capturing an image when a character recognition function is activated by the character area recognition unit;

Characterized in that it comprises a character recognition unit for recognizing the character of the image captured by the image capture unit.

Preferably, the character area recognition unit compares whether a central portion of a plurality of continuous frames is the same area for a predetermined frame, and checks whether the screen is fixed.

Preferably, the text area recognition unit includes a storage unit for storing a predetermined area of a central portion of a plurality of continuous frames, a difference value detector for extracting a predetermined area stored in the storage unit, and obtaining a difference between adjacent frames, and the difference. And comparing the sum of the difference values obtained by the value detector with a preset threshold to notify whether or not the screen is fixed.

Preferably, the character area recognition unit is configured to recognize that the camera is in close-up focal length when the number of high frequency components of a specific area of each image is more than a predetermined number.

Character recognition method in the camera phone according to an embodiment of the present invention,

Receiving a series of constant images captured by a camera; Confirming whether the preview screen is fixed by comparing specific regions of the continuous constant images; If the preview result is fixed, the method further includes determining whether the focus of the continuous image is greater than or equal to a threshold for notification of focus of the image, and automatically performing character recognition when the preview screen is greater than or equal to the threshold.

Preferably, the fixing of the preview screen comprises: extracting a predetermined region of a frame center portion of a continuous image, comparing each of the extracted regions with a previous frame, and obtaining respective difference values; Obtaining a sum of the sums, comparing the obtained sum value with a predetermined threshold value, and indicating that the sum is the same area.

Preferably, the focus confirmation of each image is characterized in that it is determined that the macro focus distance is when the number of high frequency components of a specific region of each image is a predetermined number or more.

Preferably, after storing the focus state of the previous frames continuously, if the successive values are all within the focal length, it is recognized as a text area that is in focus.

Hereinafter, the present invention will be described with reference to the accompanying drawings. 2 is a block diagram showing a character recognition device in the camera phone according to the present invention.

Referring to FIG. 2, an image capturing unit 10 for capturing an image, a character region recognizing unit 20 which indicates whether a predetermined frame is the same region and a focal length of the photographed image are well matched; The image capturing unit 30 captures an image for character recognition by the character area recognizing unit 20, and a character recognition unit 40 recognizing a character in the captured image area.

Referring to the accompanying drawings, a character recognition apparatus for a camera phone according to an embodiment of the present invention as described above is as follows.

Referring to FIG. 2, the image capturing unit 10 includes a camera imaging device and captures an image input from a subject. The text area recognition unit 20 informs the image capturing unit 30 whether the text area and the frame desired by the user are between the image capturing unit 10 and the image capturing unit 30.

To this end, if the user wants to recognize a character, the character stops at the center in the preview screen. In this case, the text area recognition unit 20 may indicate whether the same area is applied to a plurality of consecutive frames inputted from the image capturing unit 10. That is, it is determined whether the same image is the same for the consecutive images input under the condition that the camera position is not moved while the user focuses. Here, the image for region determination is not a single frame but a series of N or more (N> 1) frames.

In detail, the character area recognizer 20 may include an image frame storage 21, a difference value detector 22 that detects a difference between two adjacent frames, and an adder that obtains a sum of the difference values as illustrated in FIG. 3. 23) and a comparison unit 24 which compares the value summed up by the adding unit 23 with a threshold and notifies the same area (area fixed notification) when the value is less than or equal to the threshold.

The image frame storage unit 21 stores a predetermined area of the M frames f1 to fm, and the difference value detection units 22: 22a to 22n are a constant of the center part fc of each frame f1 to fm. The difference between the value extracted from the region and the value extracted from the same region of the previous frame is respectively detected. That is, the difference value detector 22 detects the difference between two adjacent frames, and thus the N difference value detectors 22 (22a to 22n) are output. Here, the image frame storage 21 may store only a central region fc or a specific region of the central region fc for detecting a difference value without storing one complete frame. In this way, the calculation amount can be reduced by acquiring and calculating only a specific area of the center portion of the image.

The adder 23 adds the difference values detected by the difference value detector 22, and the comparator 24 compares the summed value with a preset threshold and then, if the comparison result is equal to or less than the threshold, the same area. That is, it is output to the screen capture unit 30 as the area fixed notification information.

In this way, the text area recognition unit 20 transmits to the screen capture unit 30 that the preview screen is fixed. This may be a function of the fixed screen notification algorithm.

In addition, the text area recognition unit 20 performs the focus notification function while confirming that the preview screen is fixed.

The focus notification function in the text area recognition unit 20 applies a method of obtaining a high frequency component using a Laplacian operator mask for each image. For example, a threshold is determined by limiting a specific situation, and the sum of difference values in horizontal, vertical, and diagonal directions is compared with the threshold as shown in FIG. 4. Here, when the average of the difference is obtained, a large change value in a specific direction may represent a value lower than the average. The threshold value is set to a range value of a common change area that can be obtained through experiments.

In general, the threshold for distinguishing high-frequency components of a specific area of the image is variable, so the value varies depending on the situation.However, a camera that acquires a character and photographs the character has a certain length of affix range. If this is possible, it is possible because experimental data can obtain a common high frequency component for the character region. The macro is generally referred to as a macro or close-up of a photo taken within 1 meter, and when a small object or a small part is taken, the camera is taken close to the subject and taken large.

5 and 6, the well-focused character image has a high frequency component in a specific range, so that when the number of components in the region is greater than or equal to a predetermined value, it may be numerically understood that the macro image is in the macro focal length of the camera.

FIG. 5 is a graph illustrating a comparison of frequency distributions of captured images. FIG. 6 is a diagram illustrating focus comparison of an image. F0 is a well-focused image. The image was shot farther than the focus, and F0-2 was shot farther than the second. Here, it can be seen that the average number of frequencies compared to the difference value level in the A1 region in the focused image F0 is high.

It is possible to obtain a clear image by numerically calculating whether each image is well focused. However, in order to automatically perform the operation between the image capturing unit 30 and the character recognizing unit 40, it is difficult to confirm the focus of a single frame.

To this end, it is difficult to determine only one frame of the image coming into the camera in order to check whether the character desired by the user is in the center of the preview screen. Accordingly, after storing the focus state of successive previous frames, when the successive focus state values are all within the focal length, the image capture unit 30 is notified of the well-captured text area.

As such, when the user moves the position of the camera in a focused state through the screen fixation notification and the focus notification function of the text area recognition unit 20, it may be prevented from switching to the automatic text recognition state.

As such, when both the screen fixation and focus notification functions are activated, the text area recognition unit 20 transmits an activation signal to the image capture unit 30. The image capture unit 30 captures the captured image based on the screen fixation and focus notification function, and the character recognition unit 40 recognizes the captured image as a character. Since the screen is fixed and receives a focused image, the recognition rate can be relatively improved.

7 is a flowchart showing a character recognition method in the camera phone according to the present invention.

Referring to FIG. 7, when an image is input (S51), it is determined whether a predetermined area in the center of a predetermined frame is the same area using a screen fixing notification algorithm (S52), and whether the preview screen is fixed by the screen fixing notification algorithm. It is confirmed (S53).

In this case, the screen freeze notification algorithm detects whether a central portion of a plurality of consecutive frames is the same area, and determines whether or not the preview screen is fixed according to the detection result. At this time, if it is not the same area, it is determined that the screen is not fixed and moves to step S51. If the sum of the difference values between two adjacent frames is less than or equal to the threshold, the screen is recognized as a fixed state.

Thereafter, if the preview screen is fixed, focus notification is performed through a focus notification algorithm (S54). That is, when the number of high frequency components in a specific range of consecutive images is more than a certain number, it is determined that it is within the macro focal length of the camera.

At this time, it is checked whether the focus of the continuous image is greater than or equal to a preset threshold value (S55), and if the check result is greater than or equal to the threshold value, the image is recognized as a well-focused image, and automatically goes to the character recognition step. The character is automatically recognized (S56).

As such, by automatically moving to the character recognition step through the fixed screen notification function and the focus notification function, the image is captured based on this, and the characters of the image are recognized.

So far, the present invention has been described with reference to the preferred embodiments, and those skilled in the art to which the present invention pertains to the detailed description of the present invention and other forms of embodiments within the essential technical scope of the present invention. Could be implemented. Here, the essential technical scope of the present invention is shown in the claims, and all differences within the equivalent range will be construed as being included in the present invention.

As described above, the present invention provides an apparatus and method for character recognition in a camera phone, which automatically recognizes a character in a captured image through a focus notification and a fixed screen notification function, thereby enabling character recognition.

In addition, since the present invention obtains and calculates only a specific area of the center portion of each frame, the calculation amount is relatively small, and since only the focused image is input, the recognition rate is relatively high.

In addition, since it automatically moves to the character recognition step, there is an advantage that can be applied to the camera character recognition without the inconvenience of the user manually pressing the key button, without mechanical automatic focusing function.

Claims (8)

In the character recognition device of the camera phone, An image capturing unit for capturing an image; A text area recognition unit for activating an automatic text recognition function after checking whether a specific area of a plurality of continuous frames inputted from the image capturing unit is matched and a focus state of the continuous predetermined frames; An image capture unit for capturing an image when a character recognition function is activated by the character area recognition unit; And a character recognition unit for recognizing a character of an image captured by the image capturing unit. The method of claim 1, And the character area recognition unit compares whether a central portion of a plurality of continuous frames is the same area for a predetermined frame and checks whether the screen is fixed. The method of claim 1, The character area recognition unit may include a storage unit for storing a predetermined area of a central portion of a plurality of continuous frames, a difference value detection unit for extracting a predetermined area stored in the storage unit, and obtaining a difference between adjacent frames, and the difference value detection unit; And a comparison unit which notifies whether or not the screen is fixed by comparing the sum of the difference values obtained by the preset threshold value. The method of claim 1, And the character area recognition unit recognizes that the macro focal length of the camera is at a close focal length when the number of high frequency components of a specific area of each image is more than a predetermined number. Receiving a series of constant images captured by a camera; Confirming whether the preview screen is fixed by comparing specific regions of the continuous constant images; Confirming a focus of the series of predetermined images; And And when the preview screen is fixed and the focus is confirmed, checking whether the focus of the continuous image is greater than or equal to a threshold for notification of focus of the image, and automatically performing character recognition if the focus is greater than or equal to a threshold. Character recognition method in a camera phone characterized by the above-mentioned. The method of claim 5, The fixing of the preview screen may include extracting a predetermined region of a center portion of a frame of consecutive images, comparing the extracted region with a previous frame to obtain respective difference values, and adding the sum of the difference values. Obtaining, comparing the obtained sum value with a predetermined threshold value, and notifying the same area if less than the threshold value character recognition method in a camera phone. The method of claim 5, Checking the focus of each image is a character recognition method for a camera phone, characterized in that it is determined that the macro focus distance when the number of high-frequency components of a specific region of each image is a predetermined number or more. The method of claim 7, wherein And after storing the focus state of the previous frames successively, if the successive values are all within the focal length, the character recognition method of the camera phone characterized in that the character area is in focus.

Images