CN110032996B - Character inclination correcting device and method based on classification - Google Patents
Character inclination correcting device and method based on classification Download PDFInfo
- Publication number
- CN110032996B CN110032996B CN201810026895.4A CN201810026895A CN110032996B CN 110032996 B CN110032996 B CN 110032996B CN 201810026895 A CN201810026895 A CN 201810026895A CN 110032996 B CN110032996 B CN 110032996B
- Authority
- CN
- China
- Prior art keywords
- character
- target objects
- category
- image
- classification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/146—Aligning or centring of the image pick-up or image-field
- G06V30/1475—Inclination or skew detection or correction of characters or of image to be recognised
- G06V30/1478—Inclination or skew detection or correction of characters or of image to be recognised of characters or characters lines
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Character Input (AREA)
Abstract
The invention provides a character inclination correction method based on classification, which comprises the following steps: obtaining an input character image; identifying a plurality of target objects from the input character image; classifying the plurality of target objects into at least one candidate character category; determining a dominant character type from the at least one candidate character type; calculating a corresponding inclination angle of each target object in the main character category; and performing image tilt correction on the input image according to the calculated tilt angle of the main character type to generate a corrected image.
Description
Technical Field
The present invention relates to image processing, and more particularly, to a classification-based character tilt correction apparatus and method.
Background
In visual image detection, characters (characters/characters) in an input image to be recognized are inclined due to printing or positioning. In addition, in practical applications, many character tilts in different situations, such as different fonts (dot matrix characters, printed forms), different sizes of fonts, disordered backgrounds, etc., may affect the accuracy of character recognition. Therefore, before character recognition processing is performed on characters in the image, it is necessary to detect the character direction and correct the character direction.
In conventional tilt correction algorithms, the overall tilt variation of the character in the image to be recognized is usually considered directly, such as Hough Transform (Hough Transform), radon Transform (random Transform), Minimum Standard Deviation (Minimum Standard development), and shortest distance flat (shortest distance flat) method. In other words, the conventional tilt correction algorithms search the angle of the best solution of the whole characters in the input image to determine the tilt angle, and cannot consider the relationship between individual characters, so that the tilt angles obtained by these methods are unstable and are easily affected by other factors (such as different fonts, mixed capital and small cases of characters, and noise interference), thereby causing a character correction error.
Therefore, a classification-based character tilt correction apparatus and a method thereof are needed to solve the above problems.
Disclosure of Invention
The invention provides a character inclination correction method based on classification, which comprises the following steps: obtaining an input character image; identifying a plurality of target objects from the input character image; classifying the plurality of target objects into at least one candidate character category; determining a dominant character type from the at least one candidate character type; calculating a corresponding inclination angle of each target object in the main character category; and performing image tilt correction on the input image according to the calculated tilt angle of the main character type to generate a corrected image.
The present invention further provides a character tilt correction apparatus of the classification base type, comprising: a memory unit for storing a character tilt correction program; and a processing unit for reading from the memory unit and executing the character tilt correction program to perform the steps of: obtaining an input character image; identifying a plurality of target objects from the input character image; classifying the plurality of target objects into at least one candidate character category; determining a dominant character type from the at least one candidate character type; calculating a corresponding inclination angle of each target object in the main character category; and performing image tilt correction on the input image according to the calculated tilt angle of the main character type to generate a corrected image.
Drawings
FIG. 1 is a block diagram illustrating an exemplary character tilt correction apparatus according to the present invention.
FIG. 2A is a schematic diagram illustrating an embodiment of calculating an object height of a target object.
FIG. 2B is a schematic diagram illustrating an object height of a target object according to another embodiment of the invention.
FIG. 3 is a flowchart illustrating a classification-based character tilt correction method according to an embodiment of the present invention.
FIGS. 4A-4F are schematic diagrams illustrating classification conditions for classifying target objects according to an embodiment of the present invention.
Description of reference numerals:
100-character tilt correction means;
110 to a processing unit;
120 to memory cells;
121-volatile memory;
122 — nonvolatile memory;
123-character tilt correction procedure;
210-inputting a character image;
θ1-rotation angle;
YTP1、YBP1、YTP2、YBP2-coordinates;
S310-S360.
Detailed Description
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
FIG. 1 is a block diagram illustrating an exemplary character tilt correction apparatus according to the present invention. As shown in FIG. 1, the character slant correction apparatus 100 includes a processing unit 110 and a memory unit 120. The memory unit 120 includes a volatile memory 121 and a non-volatile memory 122. The volatile memory 121 may be a random access memory, such as a Static Random Access Memory (SRAM) or a Dynamic Random Access Memory (DRAM), but the invention is not limited thereto. The non-volatile memory 122 may be, for example, a hard disk drive, a solid state drive, a flash memory, or a read-only memory, but the invention is not limited thereto.
The non-volatile memory 122 stores a character tilt correction program 123, and the processing unit 110 reads the character tilt correction program 123 from the non-volatile memory 122 to the volatile memory 121 and executes the character tilt correction program, wherein the character tilt correction program 123 includes a program code of a character tilt correction method.
The character tilt correction method of the present invention can consider the direction change of a part of characters in the input character image, and does not directly consider the direction change of the whole of all characters. Since the overall direction change of all characters is easily affected by noise, the difference of the characters themselves affects the calculation result. For example, there is a string "A" - "in the input character image, where A is clearly different from other words. When calculating the tilt angle of the string, the relationship of the character a may cause the overall direction of the characters of the entire string to shift, which results in an error of the calculated tilt angle. In contrast, if only the change in the direction of the character "is considered, the calculated tilt angle will be correct.
FIG. 2A is a schematic diagram illustrating an embodiment of calculating an object height of a target object. FIG. 2B is a schematic diagram illustrating an object height of a target object according to another embodiment of the invention. In the present embodiment, the target object is exemplified by a character, such as "D" (capital letters), "x" (lowercase letters), "or; "punctuation mark.
As shown in fig. 2A, the target object may be, for example, an upper case english letter D. Assuming that the rotation angle of the target object is 0, the processing unit 110 may calculate a vertex coordinate TP and a base point coordinate BP of the target object, wherein the vertex refers to a highest coordinate point of the target object in the vertical direction (Y axis), and the base point refers to a lowest coordinate of the target object in the vertical direction (Y axis). It should be noted that when the input character image rotates with its center point, the vertex and the base point of the target object are not always in fixed positions, and at different rotation angles, the vertex refers to the highest coordinate point of the target object in the vertical direction (Y axis) at the rotation angle, and the base point refers to the lowest coordinate of the target object in the vertical direction (Y axis) at the rotation angle. The embodiment of FIG. 2A is illustrated with one of the target objects in the input character image. If the number of the target objects is equal to 1, the inclination angle of the target objects can be directly determined without additional classification determination.
If the number of the target objects is greater than 1, the processing unit 110 calculates the vertex coordinates TP (X) of each rotation angle θ (for example, the vertex coordinates may be sampled at a predetermined interval, for example, 1 degree or 5 degrees) of each target object within the rotation angle range (for example, the rotation angle is between plus or minus 45 degrees)TP,YTP) And base point coordinates BP (X)BP,YBP). Wherein the vertical distance between the vertex and the base point is the object height h. Then, if the number of target objects is greater than 1 (i.e. there are multiple target objects), the processing units 110 are grouped togetherIn a combined manner, the object height difference diff (for example, only the Y-axis coordinate is considered) between any two target objects in the input character image at each rotation angle θ can be expressed by the following equation:
diffij＝|objecti(YTP)-objectj(YTP)|+|objecti(YBP)-objectj(YBP)| (1)
wherein i and j are positive integers between 1 and N, and i is not equal to j. In one embodiment, if the number of target objects is greater than 1 (i.e. there are multiple target objects), the processing unit 110 classifies any two target objects at each rotation angle according to the calculated object height difference diff between the two target objects. For example, the processing unit 110 determines whether the object height difference diff is smaller than a predetermined value, and if the object height difference is smaller than a predetermined value, the processing unit 110 may determine that the two target objects belong to the same candidate character category. And if the object height difference is not smaller than the preset value, dividing the two selected target objects into different candidate character categories.
For example, if there are 5 target objects in the input character image 210: "aDcDc" when the input character image is rotated (e.g., clockwise rotation by θ)1Angle), the 5 target objects are correspondingly rotated, as shown in fig. 2B. The total times of the comparison operation of the height difference values of the two target objects is
I.e. 10 times. The comparison between two target objects is (1) aD (2) ac (3) aD (4) ac (5) Dc (6) DD (7) Dc (8) cD (9) cc (10) Dc. That is, the height difference diff between every two target objects is compared for all the objects in different combinations. For example, for the left capital letter D, the coordinates of its vertex and end point in the vertical direction (Y-axis) are Y, respectivelyTP1And YBP1(ii) a For the capital letter D on the right, the coordinates of the vertex and the end point in the vertical direction (Y axis) are YTP2And YBP2. Is rotated by an angle theta1Corresponding height difference diff ═ YTP1-YTP2|+|YBP1-YBP2L. In this example, the objects D, D are classified into the same category, and the objects a, c are also classified into the same category.
In detail, if the characters have the same font, there are usually similar object heights, such as capital letters, and these letters have similar object height differences in several specific angles when the above determination is performed, so that the letters are classified into the same category. However, the height of the English lowercase letters is not the same, and the same letter, the same general lowercase letter (e.g. w, n, c, e, a, etc.), or the same special lowercase letter (e.g. g, j, p, q, y, etc., or f, h, b, d, k, etc.) are classified into the same category. In addition, the heights of the punctuations (+ - /) are not uniform, so that the same punctuations are classified into the same category, and different punctuations are not necessarily classified into the same category.
For example, if a string in the input character image is "WWWxxyy" -, where W is capital English letters and x and y are lowercase English letters, then it is a punctuation mark. According to the above-mentioned determination mechanism, 4 candidate character categories are obtained, such as (1) WWW, (2) xx, (3) yy, and (4) -.
In one embodiment, the processing unit 110 uses the candidate character class with the largest number of target objects as the main character class to calculate the tilt angle, for example, the calculation condition is the rotation angle corresponding to the minimum object height difference between the target objects in the main character class, which is the tilt angle Φ, which can be expressed by the following formula:
after determining the tilt angle of the main character type, the processing unit 110 performs image tilt correction on the input character image according to the determined tilt angle. For the Optical Character Recognition (OCR) process, after image tilt correction is performed on an input Character image, Character segmentation and Character Recognition are performed, and a final Character Recognition result is output.
The character inclination correction method can enable characters in the corrected image of the image inclination correction to better accord with the actual character inclination condition, and further can increase the accuracy rate of character recognition.
In some embodiments, the processing unit 110 may also use other algorithms to classify the target objects in the input character image, such as: (1) the aspect ratio of the character itself, as shown in FIG. 4A; (2) the center point of the circumscribed rectangle of the character, as shown in FIG. 4B; (3) the area of the circumscribed rectangle of the character, as shown in FIG. 4C; (4) the center of the circumscribed circle of the character, as shown in FIG. 4D; (5) the circumscribed circle area of the character, as shown in FIG. 4E; (6) the radius of the circumscribed circle of the character, as shown in FIG. 4F. That is, the processing unit 110 may classify each target object in the input character image using at least one method to generate at least one candidate character classification.
In other embodiments, if the type of the character to be recognized can be determined first, a conventional classifier can be used for classification. That is, each character to be recognized may be input into the classifier one by one for training, so as to generate a classification result of each character after training. For example, the classifier may use a Support Vector Machine (SVM), a nearest neighbor classification (K-neighbor classification), a Convolutional Neural network (Convolutional Neural Networks), or a Deep learning (Deep learning), but the present invention is not limited thereto.
In the above embodiment, when the processing unit 110 determines the dominant character type from the at least one candidate character type, in addition to the candidate character type having the minimum sum of differences in the above embodiment, the dominant character type may be determined from the at least one candidate character type according to a plurality of methods, such as: (1) has the largest sum of character areas; (2) has the greatest sum of edge strengths; (3) has the most number of characters; (4) have minimal character differences; (5) the sum of standard deviations with the largest characters; (6) the sum of the variance numbers with the largest character; (7) the average brightness of the characters is minimum (black words)/maximum (white words), but the present invention is not limited thereto.
FIG. 3 is a flowchart illustrating a classification-based character tilt correction method according to an embodiment of the present invention. In step S310, an input character image is obtained. The input character image can be from an external image capturing device or can be obtained by other computer equipment through a wired or wireless transmission interface.
In step S320, a plurality of target objects are identified from the input character image. The target object includes a character or noise. For example, before identifying the target object, the processing unit 110 may employ at least one algorithm to remove noise from the input character image, such as a threshold setting, a filter, and/or a morphology (morphology) operation. The algorithm for setting the threshold value may use the following conditions, for example: gray value intensity (gray value), edge intensity (texture), area size, height, width, or a combination thereof, but the present invention is not limited thereto. Algorithms using filters may, for example, remove noise using the following filters, for example: an average filter (mean filter), a median filter (median filter), a minimum/maximum filter (min/max filter), a peak and valley filter (peak and valley filter), a Gaussian filter (Gaussian filter), a low-pass filter (low-pass filter), a high-pass filter (high-pass filter), and the like, but the present invention is not limited thereto. The algorithm for removing noise by using morphological operations includes, for example: erosion (erosion), open computing (exposure), etc., but the invention is not limited thereto. After the noise removal process, the processing unit 110 identifies a plurality of target objects from the input character image with noise removed.
In step S330, the target objects are classified into at least one candidate character class. For example, the processing unit 110 may classify the target objects into at least one candidate character category according to a first determination mechanism, wherein the first determination mechanism may use different algorithms to classify each target object in the input character image, in addition to using the object height difference to classify the target object, such as: (1) the aspect ratio of the character itself; (2) the center point of the circumscribed rectangle of the character; (3) the area of the circumscribed rectangle of the character; (4) the center of a circumscribed circle of the character; (5) the area of the circumscribed circle of the character; (6) the radius of the circumscribed circle of the character. In other embodiments, the processing unit 110 may also use a Support Vector Machine (SVM), a nearest neighbor classification (K-neighbor classification), a Convolutional Neural network (Convolutional Neural Networks), or Deep learning (Deep learning) to pre-train and classify various possible input characters.
In step S340, a dominant character category is determined from the at least one candidate character category. For example, the processing unit 110 may determine the dominant character type from the at least one candidate character type according to a second determination mechanism, wherein the second determination mechanism may determine the dominant character type by using various methods besides the candidate character type with the minimum sum of differences to determine the dominant character type, such as: for example, in the candidate character category: (1) has the largest sum of character areas; (2) has the greatest sum of edge strengths; (3) has the most number of characters; (4) have minimal character differences; (5) the sum of standard deviations with the largest characters; (6) the sum of the variance numbers with the largest character; (7) the average brightness of the characters is minimum (black words)/maximum (white words), but the present invention is not limited thereto.
In step S350, a tilt angle corresponding to each target object in the main character category is calculated. In short, the tilt angle of each target object in the input character image is determined by the tilt angle of each target object in the main character category.
In step S360, an image tilt correction is performed on the input image according to the calculated tilt angle of the main character type to generate a corrected image. Because the method for classifying the target object and determining the main character type from the candidate character types can accurately calculate the inclination angle, the image inclination correction can be performed according to the calculated inclination angle to obtain a better effect, and the accuracy of subsequent character recognition can be improved.
In summary, the present invention provides a classification-based character tilt correction apparatus and method thereof, which can classify characters in an input character image, determine a main character type to calculate a corresponding tilt angle, and perform image tilt correction using the calculated tilt angle, so as to obtain a better corrected image. In addition, when a better corrected image is obtained, the accuracy of character segmentation and character recognition in the subsequent flow of optical character recognition can be increased.
The methods of the present invention, or certain aspects or portions thereof, may take the form of a program code (i.e., executable instructions) embodied in tangible media, such as floppy diskettes, CD-ROMS, hard drives, or any other machine-readable storage medium, wherein, when the program code is loaded into and executed by a machine, such as a computer, the machine becomes an apparatus or system for practicing the invention. The methods, systems, and apparatus of the present invention may also be embodied in the form of program code transmitted over some transmission medium, such as electrical wiring or cabling, through fiber optics, or via any other form of transmission, wherein, when the program code is received and loaded into and executed by a machine, such as a computer, the machine becomes an apparatus or system for practicing the invention. When implemented on a general-purpose processor, the program code combines with the processor to provide a unique apparatus that operates analogously to specific logic circuits.
Although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes and modifications can be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (10)
1. A classification-based character tilt correction method includes:
obtaining an input character image;
identifying a plurality of target objects from the input character image;
classifying the plurality of target objects into at least one candidate character category;
determining a dominant character type from the at least one candidate character type;
calculating a corresponding inclination angle of each target object in the main character category; and
performing image tilt correction on the input image according to the calculated tilt angle of the main character type to generate a corrected image;
wherein classifying the plurality of target objects into at least one candidate character category comprises:
calculating a vertex coordinate and a base point coordinate in each target object;
calculating an object height difference value of each rotation angle of any two target objects in the plurality of target objects in a rotation angle range; and
when the object height difference is smaller than a preset value, dividing the two calculated target objects into the same candidate character category;
and when the object height difference is not less than the preset value, dividing the two calculated target objects into different candidate character categories.
2. The classification-based character inclination correction method as claimed in claim 1, further comprising:
the candidate character category having a largest number of characters in the at least one candidate character category is determined to be the dominant character category.
3. The classification-based character inclination correction method as claimed in claim 1, further comprising:
classifying the plurality of target objects into the at least one candidate character category according to one or a combination of the aspect ratio of each target object, the center point of the circumscribed rectangle, the area of the circumscribed rectangle, the center point of the circumscribed circle, the area of the circumscribed circle, the radius of the circumscribed circle.
4. The classification-based character inclination correction method as claimed in claim 1, further comprising:
classifying the plurality of target objects into the at least one candidate character category using a support vector machine, a nearest neighbor classification method, a convolutional neural network, or deep learning.
5. The classification-based character inclination correction method as claimed in claim 1, further comprising:
the candidate character category with the largest sum of character areas, the largest sum of edge intensities, the largest number of characters, the smallest character differences, the largest sum of standard deviations of characters, the largest sum of variance of characters, or the smallest average brightness is determined as the main character category.
6. A character inclination correcting apparatus of a classification base type, comprising:
a memory unit for storing a character tilt correction program; and
a processing unit for reading from the memory unit and executing the character tilt correction procedure to perform the following steps:
obtaining an input character image;
identifying a plurality of target objects from the input character image;
classifying the plurality of target objects into at least one candidate character category;
determining a dominant character type from the at least one candidate character type;
calculating a corresponding inclination angle of each target object in the main character category; and
performing image tilt correction on the input image according to the calculated tilt angle of the main character type to generate a corrected image;
wherein the processing unit calculates a vertex coordinate and a base coordinate in each target object, calculates an object height difference value of each rotation angle of any two target objects in the plurality of target objects within a rotation angle range,
when the object height difference is smaller than a preset value, the processing unit divides the two target objects into the same candidate character category;
when the object height difference is not less than the predetermined value, the processing unit classifies the two object objects into different candidate character categories.
7. The device for classifying base character tilt correction according to claim 6, wherein said processing unit determines a candidate character category having a largest number of characters in said at least one candidate character category as said dominant character category.
8. The character slant correction device according to claim 6, wherein the processing unit classifies the plurality of target objects into the at least one candidate character category according to one or a combination of an aspect ratio of each target object, a center point of a circumscribed rectangle, an area of a circumscribed rectangle, a center point of a circumscribed circle, an area of a circumscribed circle, a radius of a circumscribed circle.
9. The classification-based character tilt correction apparatus of claim 6, wherein the processing unit classifies the plurality of target objects into the at least one candidate character class using a support vector machine, a nearest neighbor classification method, a convolutional neural network, or deep learning.
10. The apparatus of claim 6, wherein the processing unit determines the candidate character type having the largest sum of character areas, the largest sum of edge intensities, the largest number of characters, the smallest character difference, the largest sum of standard deviations of characters, the largest sum of variance of characters, or the smallest average brightness as the dominant character type.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810026895.4A CN110032996B (en) | 2018-01-11 | 2018-01-11 | Character inclination correcting device and method based on classification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810026895.4A CN110032996B (en) | 2018-01-11 | 2018-01-11 | Character inclination correcting device and method based on classification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110032996A CN110032996A (en) | 2019-07-19 |
| CN110032996B true CN110032996B (en) | 2021-06-04 |
Family
ID=67234690
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810026895.4A Active CN110032996B (en) | 2018-01-11 | 2018-01-11 | Character inclination correcting device and method based on classification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110032996B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101727581A (en) * | 2009-12-10 | 2010-06-09 | 上海名图软件有限公司 | Plate number tilt correcting method based on character pre-cut |
| CN101887521A (en) * | 2010-06-22 | 2010-11-17 | 中兴通讯股份有限公司 | Method and terminal for rectifying deviation of file |
| CN102760233A (en) * | 2011-04-25 | 2012-10-31 | 夏普株式会社 | Image processing apparatus |
| CN102938062A (en) * | 2012-10-16 | 2013-02-20 | 山东山大鸥玛软件有限公司 | Document image slant angle estimation method based on content |
| CN103400130A (en) * | 2013-07-22 | 2013-11-20 | 哈尔滨工业大学 | Energy minimization framework-based document image tilt detection and correction method |
| CN105654082A (en) * | 2014-11-12 | 2016-06-08 | 佳能株式会社 | Method and equipment for character identification post-processing and image picking equipment comprising equipment |
| CN105760901A (en) * | 2016-01-27 | 2016-07-13 | 南开大学 | Automatic language identification method for multilingual skew document image |
| CN106096601A (en) * | 2016-06-06 | 2016-11-09 | 深圳辰通智能股份有限公司 | The method and system of character types in a kind of automatic detection bill |
| CN106295638A (en) * | 2016-07-29 | 2017-01-04 | 北京小米移动软件有限公司 | Certificate image sloped correcting method and device |
| JP2017120455A (en) * | 2015-12-28 | 2017-07-06 | キヤノン株式会社 | Information processing device, program and control method |
-
2018
- 2018-01-11 CN CN201810026895.4A patent/CN110032996B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101727581A (en) * | 2009-12-10 | 2010-06-09 | 上海名图软件有限公司 | Plate number tilt correcting method based on character pre-cut |
| CN101887521A (en) * | 2010-06-22 | 2010-11-17 | 中兴通讯股份有限公司 | Method and terminal for rectifying deviation of file |
| CN102760233A (en) * | 2011-04-25 | 2012-10-31 | 夏普株式会社 | Image processing apparatus |
| CN102938062A (en) * | 2012-10-16 | 2013-02-20 | 山东山大鸥玛软件有限公司 | Document image slant angle estimation method based on content |
| CN103400130A (en) * | 2013-07-22 | 2013-11-20 | 哈尔滨工业大学 | Energy minimization framework-based document image tilt detection and correction method |
| CN105654082A (en) * | 2014-11-12 | 2016-06-08 | 佳能株式会社 | Method and equipment for character identification post-processing and image picking equipment comprising equipment |
| JP2017120455A (en) * | 2015-12-28 | 2017-07-06 | キヤノン株式会社 | Information processing device, program and control method |
| CN105760901A (en) * | 2016-01-27 | 2016-07-13 | 南开大学 | Automatic language identification method for multilingual skew document image |
| CN106096601A (en) * | 2016-06-06 | 2016-11-09 | 深圳辰通智能股份有限公司 | The method and system of character types in a kind of automatic detection bill |
| CN106295638A (en) * | 2016-07-29 | 2017-01-04 | 北京小米移动软件有限公司 | Certificate image sloped correcting method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110032996A (en) | 2019-07-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109086714B (en) | Form recognition method, recognition system and computer device | |
| CN106446896B (en) | Character segmentation method and device and electronic equipment | |
| CN107609549B (en) | Text detection method for certificate image in natural scene | |
| CN111951290B (en) | Edge detection method and device for object in image | |
| US8965127B2 (en) | Method for segmenting text words in document images | |
| US8590794B2 (en) | Barcode recognion method and computer product thereof | |
| US9104940B2 (en) | Line segmentation method applicable to document images containing handwriting and printed text characters or skewed text lines | |
| WO2021233266A1 (en) | Edge detection method and apparatus, and electronic device and storage medium | |
| WO2016033710A1 (en) | Scene text detection system and method | |
| US10318803B1 (en) | Text line segmentation method | |
| US10438083B1 (en) | Method and system for processing candidate strings generated by an optical character recognition process | |
| US11526708B2 (en) | Information processing device, information processing method, and recording medium | |
| JP6170860B2 (en) | Character recognition device and identification function generation method | |
| CN110717492A (en) | Method for correcting direction of character string in drawing based on joint features | |
| WO2021190155A1 (en) | Method and apparatus for identifying spaces in text lines, electronic device and storage medium | |
| CN112699704B (en) | Method, device, equipment and storage device for detecting bar code | |
| WO2019232870A1 (en) | Method for acquiring handwritten character training sample, apparatus, computer device, and storage medium | |
| CN116597466A (en) | Engineering drawing text detection and recognition method and system based on improved YOLOv5s | |
| RU2453919C1 (en) | Method of detecting spam in bitmap image | |
| CN110321887B (en) | Document image processing method, document image processing apparatus, and storage medium | |
| CN117612179A (en) | Method and device for recognizing characters in image, electronic equipment and storage medium | |
| CN116434071B (en) | Determination method, determination device, equipment and medium for normalized building mask | |
| CN110032996B (en) | Character inclination correcting device and method based on classification | |
| TWI766930B (en) | Classification-based character skew correction apparatus and method thereof | |
| CN109961063B (en) | Text detection method and device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |