CN115147843A - Text recognition method and device, storage medium and computer equipment - Google Patents
Text recognition method and device, storage medium and computer equipment Download PDFInfo
- Publication number
- CN115147843A CN115147843A CN202210885587.3A CN202210885587A CN115147843A CN 115147843 A CN115147843 A CN 115147843A CN 202210885587 A CN202210885587 A CN 202210885587A CN 115147843 A CN115147843 A CN 115147843A
- Authority
- CN
- China
- Prior art keywords
- text block
- text
- point
- horizontal direction
- corner
- 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.)
- Pending
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- G06V30/18—Extraction of features or characteristics of the image
-
- 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
- G06V30/19—Recognition using electronic means
- G06V30/191—Design or setup of recognition systems or techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
- G06V30/19173—Classification techniques
Abstract
According to the text identification method, the text identification device, the storage medium and the computer equipment, before a text in an image to be identified is identified, a text block in a designated area in the image to be identified and position information of the text block can be acquired, the position information can comprise coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be identified, then a rotation point and a rotation angle for rotating the text block to the horizontal direction can be determined according to the coordinates corresponding to the four corner points of the text block, and the text block is rotated to the horizontal direction according to the rotation point and the rotation angle, so that the text block located in the horizontal direction can be identified; in the process, the direction of the text block is corrected before the text block in the image to be recognized is recognized, so that the direction of the text in the text block to be recognized is fixed, the text recognition efficiency can be improved, and the accuracy of the text recognition can be improved.
Description
Technical Field
The present application relates to the field of image processing technologies, and in particular, to a text recognition method, an apparatus, a storage medium, and a computer device.
Background
OCR (optical character recognition) refers to a process of analyzing and recognizing an image file of text data to obtain characters and layout information. I.e. the text in the image is recognized and returned in the form of text. The OCR recognition technology has the characteristics of high accuracy, strong stability, simplicity, easiness in use and the like, and is widely applied to the fields of education, medical treatment, industry and the like. For example, OCR recognition technology can be applied to a PDA (scanning gun) as a complement to barcode, two-dimensional code recognition.
At present, in the process of applying the OCR recognition technology to the PDA, since a recognition scene is not fixed, a text block in an image to be recognized may have a certain inclination, so that a direction of characters in the text block is not fixed, and finally, accuracy of a recognition result is low.
Disclosure of Invention
The present application aims to solve at least one of the above technical drawbacks, and in particular, to solve the technical drawback in the prior art that when the OCR recognition technology is applied to a PDA, since a recognition scene is not fixed, text blocks in an image to be recognized may have a certain inclination, so that directions of characters in the text blocks are not fixed, and accuracy of a final recognition result is low.
The application provides a text recognition method, which comprises the following steps:
acquiring a text block of a specified area in an image to be recognized and position information of the text block, wherein the position information of the text block comprises coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be recognized;
determining a rotation point and a rotation angle for rotating the text block to the horizontal direction according to the coordinates of the four corner points of the text block;
and rotating the text block to the horizontal direction according to the rotation point and the rotation angle, and identifying the text block in the horizontal direction.
Optionally, the determining, according to coordinates of four corner points of the text block, a rotation point for rotating the text block to a horizontal direction includes:
determining the width and the height of the text block according to the coordinates of four corner points of the text block, wherein the width of the text block is not less than the height of the text block;
traversing coordinates of four corner points of the text block, and sequencing the coordinates of the four corner points clockwise or anticlockwise according to the width and the height of the text block; among the four edge points after sequencing, the distance from a first edge point to a second edge point is consistent with the width of the text block, and the distance from the second edge point to a third edge point is consistent with the height of the text block;
and taking the first corner point of the four sorted corner points as a rotation point for rotating the text block to the horizontal direction.
Optionally, the traversing coordinates of four corner points of the text block, and performing clockwise or counterclockwise sorting on the coordinates of the four corner points according to the width and the height of the text block includes:
traversing coordinates of four edge points of the text block, and taking the edge point with the smallest sum of the abscissa and the ordinate in the four edge points of the text block as a starting point;
sorting the coordinates of the remaining corner points in the four corner points in a clockwise or anticlockwise direction from the starting point to obtain a sorting result;
determining whether the distance from the first corner point to the second corner point in the sequencing result is consistent with the width of the text block or not and whether the distance from the second corner point to the third corner point in the sequencing result is consistent with the height of the text block or not according to the width of the text block, the distance from the first corner point to the second corner point in the sequencing result and the distance from the second corner point to the third corner point in the sequencing result;
if the two edge points are inconsistent, re-ordering the coordinates of the four edge points in the ordering result according to the ordering direction in the last ordering, returning to execute the step of determining whether the distance from the first edge point to the second edge point in the ordering result is consistent with the width of the text block or not and whether the distance from the second edge point to the third edge point in the ordering result is consistent with the height of the text block or not according to the width of the text block and the distance from the first edge point to the second edge point and the distance from the second edge point to the third edge point in the ordering result are consistent with the height of the text block or not after obtaining a new ordering result;
and if the two are consistent, taking the sorting result as a final sorting result.
Optionally, the determining, according to coordinates of four corner points of the text block, a rotation angle for rotating the text block to a horizontal direction includes:
determining the broadside of the text block according to the coordinates of the four corner points of the text block;
and calculating an included angle between the wide side of the text block and the coordinate axis of the first coordinate system, and determining a rotation angle when the text block is rotated to the horizontal direction according to the included angle.
Optionally, the rotating the text block to the horizontal direction according to the rotation point and the rotation angle to obtain the text block located in the horizontal direction includes:
taking a plane which is arranged opposite to the image to be recognized as a mapping plane;
taking a point in the mapping plane, which is arranged opposite to the rotation point, as a coordinate origin, and constructing a second coordinate system according to the coordinate origin;
and performing perspective transformation on the text block to the second coordinate system according to the rotation angle by taking the coordinate origin as the coordinate starting point of the text block after the perspective transformation to obtain the text block in the horizontal direction.
Optionally, the recognizing the text block located in the horizontal direction includes:
determining whether the text in the text block in the horizontal direction is inverted;
if so, performing 180-degree rotation operation on the text block in the horizontal direction, and identifying the rotated text block;
if not, directly identifying the text block in the horizontal direction.
Optionally, the determining whether the text in the text block located in the horizontal direction is inverted includes:
inputting the text block in the horizontal direction into a pre-configured text recognition model to obtain the inversion condition of the text in the text block in the horizontal direction output by the text recognition model;
the text recognition model is obtained by training with a training text block positioned in the horizontal direction as a training sample and with the inversion condition of the text marked in the training text block as a sample label.
The present application further provides a text recognition apparatus, including:
the system comprises a text acquisition module, a recognition module and a recognition module, wherein the text acquisition module is used for acquiring a text block of a specified area in an image to be recognized and position information of the text block, and the position information of the text block comprises coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be recognized;
the rotation parameter determining module is used for determining a rotation point and a rotation angle for rotating the text block to the horizontal direction according to the coordinates of the four corner points of the text block;
and the text recognition module is used for rotating the text block to the horizontal direction according to the rotation point and the rotation angle and recognizing the text block in the horizontal direction.
The present application further provides a storage medium having stored therein computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the text recognition method as described in any one of the above embodiments.
The present application further provides a computer device, comprising: one or more processors, and a memory;
the memory has stored therein computer-readable instructions that, when executed by the one or more processors, perform the steps of the text recognition method of any of the above embodiments.
According to the technical scheme, the embodiment of the application has the following advantages:
according to the text identification method, the text identification device, the storage medium and the computer equipment, before a text in an image to be identified is identified, a text block in a designated area in the image to be identified and position information of the text block can be acquired, the position information can comprise coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be identified, then a rotation point and a rotation angle for rotating the text block to the horizontal direction can be determined according to the coordinates corresponding to the four corner points of the text block, for example, the broadside and the rotation angle of the broadside of the text block can be determined according to the coordinates corresponding to the four corner points of the text block, then any point on the broadside is taken as a rotation point, and the text block is rotated to the horizontal direction according to the rotation point and the rotation angle, so that the text block located in the horizontal direction can be identified; in the process, the direction of the text block is corrected before the text block in the image to be recognized is recognized, so that the direction of the text in the text block to be recognized is fixed, the text recognition efficiency can be improved, and the accuracy of the text recognition can be improved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
Fig. 1 is a schematic flowchart of a text recognition method according to an embodiment of the present application;
fig. 2 is a schematic structural diagram of a text block in an image to be recognized according to an embodiment of the present application;
fig. 3 is a text block located in a horizontal direction according to an embodiment of the present application;
FIG. 4 is a block of text in a horizontal orientation according to an embodiment of the present disclosure;
fig. 5 is a schematic structural diagram of a text recognition apparatus according to an embodiment of the present application;
fig. 6 is a schematic internal structural diagram of a computer device according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only some embodiments of the present application, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
At present, in the process of applying the OCR recognition technology to the PDA, since a recognition scene is not fixed, a text block in an image to be recognized may have a certain inclination, so that a direction of characters in the text block is not fixed, and finally, accuracy of a recognition result is low. Therefore, the following technical solutions are proposed in the present application, specifically see the following:
in an embodiment, as shown in fig. 1, fig. 1 is a schematic flowchart of a text recognition method provided in an embodiment of the present application; the application provides a text recognition method, which can comprise the following steps:
s110: and acquiring a text block of a specified area in the image to be recognized and position information of the text block.
In this step, when a text block in an image to be recognized is recognized, a text block in a designated area in the image to be recognized and position information of the text block may be obtained first, where the position information of the text block in the present application may include coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be recognized.
The image to be identified obtained by the application can be a text image in different scenes, such as a text image in a scene of a certificate, a commodity number and a product weight, and can also be a text image in other scenes, which is not repeated herein.
It will be appreciated that the present application provides an OCR recognition component that employs the sight positioning as a reference point and captures a number of text blocks appearing in a field of view (FOV) in an image to be recognized. Therefore, in the present application, the designated area in the image to be recognized may be an image area appearing in the FOV in the image to be recognized, where the image area includes at least one text block, each text block is a rectangular text block, the rectangular text block corresponds to four corner points, and coordinates of the corner points of each rectangular text block are different.
Specifically, when coordinates of four corner points of a text block in a specified area in an image to be recognized are obtained, a first coordinate system may be constructed according to the image to be recognized, and then coordinates of the four corner points of the text block in the specified area in the first coordinate system are determined. When the first coordinate system is constructed, one of the four corner points of the image to be recognized may be used as a coordinate origin, and the first coordinate system is constructed through the coordinate origin, or a center point of the image to be recognized may be used as the coordinate origin, and the first coordinate system is constructed through the coordinate origin, of course, points other than the above points in the image to be recognized may also be used as the coordinate origin, and the first coordinate system is constructed through the coordinate origin, which is specifically set according to actual conditions, and is not limited herein.
In a specific implementation manner, the application may use the upper left corner point of the image to be recognized as the origin of coordinates, the right direction is the X forward direction, the downward direction is the Y forward direction to give a first coordinate system, and the coordinates of four corner points of each text block in the image to be recognized in the first coordinate system are determined by using the pixel points as units. Schematically, as shown in fig. 2, fig. 2 is a schematic structural diagram of a text block in an image to be recognized according to an embodiment of the present application; in fig. 2, the upper left corner point of the image to be recognized is used as the origin of coordinates, and the first coordinate system is constructed according to the origin of coordinates, so that the coordinates corresponding to the four corner points of the text block in the first coordinate system can be obtained, which are (295, 173), (482, 124), (670, 820), (482, 870), respectively.
S120: and determining a rotation point and a rotation angle for rotating the text block to the horizontal direction according to the coordinates of the four corner points of the text block.
In this step, after the text block in the designated area in the image to be recognized and the coordinates of the four corner points of the text block in the first coordinate system are obtained through S110, the rotation point and the rotation angle for rotating the text block to the horizontal direction may be determined according to the coordinates of the four corner points of the text block, and the text block may be rotated to the horizontal direction according to the rotation point and the rotation angle.
It can be understood that, since the recognition scenes of the image to be recognized in the present application are complex and variable, the direction of the characters in the text block in the designated area in the image to be recognized is not fixed, and if the text block is directly input into the recognition model for text recognition, the efficiency and accuracy of text recognition may be affected, for example, the character recognition is mistaken due to the inclination of the direction of the characters. Therefore, after the text block in the designated area in the image to be recognized and the coordinates of the four corner points of the text block are obtained, the rotation point and the rotation angle for rotating the text block to the horizontal direction can be determined according to the coordinates of the four corner points, so that the text direction of the inclined text block can be corrected according to the rotation point and the rotation angle, and the text block input to the recognition model is ensured to be the text block located in the horizontal direction.
Further, when determining a rotation point and a rotation angle for rotating a text block to a horizontal direction according to four corner points of the text block, the method and the device for rotating a broadside of the text block to a horizontal direction may determine the broadside of the text block according to the four corner points, and then determine a rotation point for rotating one broadside of the text block to the horizontal direction according to the four corner points and a connection line between the four corner points, where the rotation point may be any one of the four corner points, any one of the connection lines between the four corner points, a center point of the text block, or the like. After the rotation point for rotating the text block to the horizontal direction is determined, the rotation angle for rotating the text block to the horizontal direction can be determined according to the coordinates of the four corner points of the text block. For example, the method and the device can calculate the inclination angle of the broadside of the text block relative to the upper edge line or the lower edge line of the image to be recognized through the coordinates of four edge points, and take the inclination angle as the rotation angle; or the inclination angle of the broadside of the text block relative to the horizontal line corresponding to the rotation point may be calculated according to the rotation point of the text block and the coordinates of the four corner points, and the inclination angle is used as the rotation angle, or other manners of calculating the rotation angle of the text block according to the coordinates of the four corner points of the text block may be used. The specific calculation method may be pythagorean theorem, rotation matrix, vector rotation, etc., which is not limited herein.
In addition, the rotation angle of the present application includes both a rotation angle value and a rotation direction, the rotation angle value may be between 0 and 180 degrees, and the rotation direction may be a clockwise direction or a counterclockwise direction.
S130: and rotating the text block to the horizontal direction according to the rotation point and the rotation angle, and identifying the text block in the horizontal direction.
In this step, after the rotation point and the rotation angle for rotating the text block to the horizontal direction are determined according to the coordinates of the four corner points of the text block in S110, the text block may be rotated to the horizontal direction according to the rotation point and the rotation angle, and the text block located in the horizontal direction is identified.
Specifically, after a rotation point and a rotation angle for rotating the text block to the horizontal direction are obtained, the broadside of the text block can be rotated to the horizontal direction by the rotation point and the rotation direction, the specific rotation operation can be performed in a spatial transformation and mapping mode, the text block in the image to be recognized is subjected to perspective transformation to the text block in the horizontal direction, then the text block is cut, and text recognition is performed on the cut text block.
In the above embodiment, before identifying a text in an image to be identified, a text block in a specified area in the image to be identified and position information of the text block may be obtained, where the position information may include coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be identified, and then a rotation point and a rotation angle for rotating the text block to a horizontal direction may be determined according to the coordinates corresponding to the four corner points of the text block, for example, a broadside and a rotation angle of the broadside of the text block may be determined according to the coordinates corresponding to the four corner points of the text block, and then any point on the broadside is taken as a rotation point, and the text block is rotated to the horizontal direction according to the rotation point and the rotation angle, thereby identifying the text block located in the horizontal direction; in the process, the direction of the text block is corrected before the text block in the image to be recognized is recognized, so that the direction of the text in the text block to be recognized is fixed, the text recognition efficiency can be improved, and the accuracy of the text recognition can be improved.
In one embodiment, the determining, in S120, rotation points for rotating the text block to a horizontal direction according to coordinates of four corner points of the text block may include:
s121: and determining the width and the height of the text block according to the coordinates of the four corner points of the text block, wherein the width of the text block is not less than the height of the text block.
In this step, when determining a rotation point for rotating the text block to the horizontal direction, the width and the height of the text block may be determined first, and then the rotation point for rotating the text block to the horizontal direction may be determined according to the coordinates of the four corner points of the text block and the width and the height of the text block.
It will be appreciated that since a typical block of text is made up of one or more characters, the width of a single block of text is typically no less than the height. Based on the method, the width and the height of the text block can be determined according to coordinates of four corner points, the calculation method can be that the width of the text block is determined according to the average value of the lengths of the upper side and the lower side of the text block, the height of the text block is determined according to the average value of the lengths of the left side and the right side of the text block, and if the width is smaller than the height, the width and the height are replaced.
It should be noted that, here, the upper side refers to a connecting line closest to an upper edge line of the image to be recognized among connecting lines of four edge points of the text block, the lower side refers to a connecting line closest to a lower edge line of the image to be recognized among connecting lines of four edge points of the text block, similarly, the left side of the text block refers to a connecting line closest to a left edge line of the image to be recognized among connecting lines of four edge points of the text block, and the right side refers to a connecting line closest to a right edge line of the image to be recognized among connecting lines of four edge points of the text block.
Taking fig. 2 as an example, the coordinates of the four corner points in fig. 2 are: (295, 173), (482, 124), (670, 820), (482, 870), the width and the height of the text block can be calculated to be respectively the width 721 and the height 194 according to the coordinates of the four corner points.
S122: traversing coordinates of four corner points of the text block, and sequencing the coordinates of the four corner points clockwise or anticlockwise according to the width and the height of the text block; among the four edge points after the sorting, the distance from the first edge point to the second edge point is consistent with the width of the text block, and the distance from the second edge point to the third edge point is consistent with the height of the text block.
In this step, after the coordinates of the four corner points of the text block are obtained, the coordinates of the four corner points can be traversed, the four coordinates are clockwise sorted according to the width and the height of the text block, in the four sorted corner points, the distance from the first corner point to the second corner point is consistent with the width of the text block, and the distance from the second corner point to the third corner point is consistent with the height of the text block, so that the subsequent rotation operation on the text block can be facilitated.
S123: and taking the first corner point of the four sorted corner points as a rotation point for rotating the text block to the horizontal direction.
In this step, after the four corner points of the text block are sequenced, because the distance from the first corner point to the second corner point is consistent with the width of the text block and the distance from the second corner point to the third corner point is consistent with the height of the text block among the sequenced four corner points, at this time, if the first corner point of the four corner points is selected as the rotation point, not only the rotation angle of the text block is easy to determine, but also the text block can be rotated to the horizontal direction more quickly.
Furthermore, the four coordinates can be sorted anticlockwise according to the width and the height of the text block, a first corner point of the sorted four corner points is used as a rotation point for rotating the text block to the horizontal direction, and the traversing directions of the coordinates in the text block before and after rotation are consistent. When the clockwise sorting or the counterclockwise sorting is selected, the determination may be performed according to a construction manner of the first coordinate system, for example, when the first coordinate system constructs a coordinate system by using an upper left corner of the image to be recognized as an origin of coordinates, and uses a right corner as an X forward direction, and uses a downward corner as a Y forward direction, at this time, a clockwise sorting manner may be selected, and when the first coordinate system constructs a coordinate system by using a lower right corner of the image to be recognized as an origin of coordinates, and uses a left corner as an X forward direction, and uses a left corner as a Y forward direction, and uses the upward corner as a Y forward direction, at this time, a counterclockwise sorting manner may be selected.
Furthermore, the second, third, or fourth corner point of the four sorted corner points can be selected as a rotation point for rotating the text block to the horizontal direction, and when different corner points are selected as rotation points, the rotation angles of the corresponding text blocks can be different.
In one embodiment, traversing coordinates of four corner points of the text block in S122, and sorting the coordinates of the four corner points clockwise or counterclockwise according to the width and height of the text block may include:
s1221: traversing coordinates of four edge points of the text block, and taking the edge point with the smallest sum of the abscissa and the ordinate in the four edge points of the text block as a starting point;
s1222: sorting the coordinates of the remaining corner points in the four corner points in a clockwise or anticlockwise direction from the starting point to obtain a sorting result;
s1223: determining whether the distance from the first corner point to the second corner point in the sequencing result is consistent with the width of the text block or not and whether the distance from the second corner point to the third corner point in the sequencing result is consistent with the height of the text block or not according to the width of the text block, the distance from the first corner point to the second corner point in the sequencing result and the distance from the second corner point to the third corner point in the sequencing result; if there is an inconsistency, execute S1224; if both are consistent, S1225 is executed.
S1224: and re-ordering the coordinates of the four corner points in the ordering result according to the ordering direction in the last ordering, obtaining a new ordering result, and returning to the step of executing S1223 until the results are uniform.
S1225: and taking the sorting result as a final sorting result.
In this embodiment, when coordinates of four corner points in a text block are sorted, a corner point with the smallest sum of an abscissa and an ordinate in the four corner points can be used as a starting point, and coordinates of remaining corner points in the four corner points are sorted clockwise or counterclockwise from the starting point to obtain a sorting result, and then, according to the width and height of the text block, a distance from a first corner point to a second corner point and a distance from the second corner point to a third corner point in the sorting result, whether the distance from the first corner point to the second corner point in the sorting result is consistent with the width of the text block or not is determined, whether the distance from the second corner point to the third corner point is consistent with the height of the text block or not is determined, if the distances are inconsistent, coordinates of the four corner points in the sorting result are reordered according to the sorting direction of the last time, and after a new sorting result is obtained, whether the distance from the first corner point to the second corner point in the sorting result is consistent with the width of the text block or not, whether the distance from the second corner point to the third corner point in the sorting result is consistent with the width of the text block or not; and if the two are consistent, taking the sorting result as a final sorting result.
In a specific implementation manner, when a first coordinate system of the application uses an upper left corner point of an image to be identified as a coordinate origin to construct a coordinate system, and uses a right corner point as an X positive direction and a downward corner point as a Y positive direction, an X coordinate + a Y coordinate of each corner point coordinate can be determined first, the corner point corresponding to the minimum value of the X coordinate + the Y coordinate is used as the upper left corner point, and the order of the coordinates of the points is determined to be clockwise or counterclockwise through the first three points in the coordinates of the current four corner points, if the coordinates are counterclockwise, the coordinates of the 2 nd and 4 th corner points are exchanged, so that the coordinates of the four corner points are ensured to be ordered clockwise; next, if the 1 st is found, the distance between the 2 nd corner points is smaller than the distance between the 2 nd and the 3 rd corner points, at this time, the order of the four corner points can be reordered according to 2,3,4,1 to update the coordinates of the points, the coordinates of the four points thus obtained are clockwise, and the upper left corner point of the text block is taken as the starting point.
Taking fig. 2 as an example, first, the first point of the four corner points in fig. 2 is determined as a (295, 173) coordinate point, the coordinates of the corner points corresponding clockwise are (482, 124), (670, 820), (482, 870), and the coordinates are corrected to be (482, 124), (670, 820), (482, 870), (295, 173) according to the width-height relationship, and at this time, the corner point corresponding to the coordinates (482, 124) can be used as a rotation point.
In one embodiment, the determining, in S120, rotation angles for rotating the text block to a horizontal direction according to coordinates of four corner points of the text block may include:
s210: and determining the broadside of the text block according to the coordinates of the four corner points of the text block.
S211: and calculating an included angle between the wide edge of the text block and the coordinate axis of the first coordinate system, and determining a rotation angle when the text block is rotated to the horizontal direction according to the included angle.
In this embodiment, when the rotation angle for rotating the text block to the horizontal direction is determined, the broadside of the text block may be determined according to coordinates of four corner points of the text block, then an included angle between the broadside of the text block and a coordinate axis of the first coordinate system is calculated, and the rotation angle for rotating the text block to the horizontal direction is determined according to the included angle.
Specifically, when the rotation angle of the text block is determined, the broadside of the text block may be optionally determined, and specifically, the determination may be performed by coordinates of four corner points of the text block. For example, the width and height of the text block can be determined according to coordinates of four corner points, and the calculation method can be that the width of the text block is determined according to the average value of the lengths of the upper side and the lower side of the text block, the height of the text block is determined according to the average value of the lengths of the left side and the right side of the text block, if the width is smaller than the height, the width of the text block and the height are replaced, and the width of the text block is determined according to the final result.
According to the method and the device, the text block is required to be rotated to the horizontal direction, namely, the broadside of the text block is rotated to the horizontal direction, so that after the broadside of the text block is determined, an included angle between the broadside and a coordinate axis of the first coordinate system can be calculated, and a rotation angle for rotating the text block to the horizontal direction is determined according to the included angle.
For example, when the first coordinate system in the present application uses the upper left corner point of the image to be recognized as the origin of coordinates to construct a coordinate system, and uses the right direction as the X forward direction and the downward direction as the Y forward direction, the present application may calculate an included angle between the wide side of the text block and the X axis of the first coordinate system, determine the rotation direction when the text block rotates to the X axis, and finally determine the rotation angle according to the included angle and the rotation direction. Of course, if the X axis of the first coordinate system is not a coordinate axis on the horizontal line, but the longitudinal axis is a coordinate axis on the horizontal line, the present application may calculate an angle between the wide side of the text block and the Y axis of the first coordinate system, determine a rotation direction when the text block rotates to the Y axis, and finally determine the rotation angle according to the angle and the rotation direction.
In one embodiment, the rotating the text block to the horizontal direction according to the rotation point and the rotation angle in S130 to obtain the text block located in the horizontal direction may include:
s131: and taking a plane which is arranged opposite to the image to be recognized as a mapping plane.
S132: and taking a point in the mapping plane, which is arranged opposite to the rotation point, as a coordinate origin, and constructing a second coordinate system according to the coordinate origin.
S133: and performing perspective transformation on the text block to the second coordinate system according to the rotation angle by taking the coordinate origin as the coordinate starting point of the text block after the perspective transformation to obtain the text block in the horizontal direction.
In this embodiment, when a text block is rotated to the horizontal direction, a plane set opposite to an image to be recognized may be selected as a mapping plane, a point set opposite to a rotation point in the mapping plane is used as a coordinate origin, a second coordinate system is constructed according to the coordinate origin, and then the coordinate origin is used as a coordinate starting point of the text block after perspective transformation, and the text block is perspective-transformed into the second coordinate system according to a corresponding rotation angle, so that the text block located in the horizontal direction is obtained.
Taking fig. 2 as an example, the application may use the upper left corner of the image to be recognized as the origin, the right side as the X forward direction, and the downward side as the Y forward direction to give the first coordinate system, and calculate the rotation angles of the two broad sides of the text block, where the range is (0, 180) degrees. The rotation angle of the text block is determined according to the rotation angle mean value of the two broadsides, and the rotation angles of the two broadsides determined in fig. 2 are both approximate to 74 degrees, so the rotation angle mean value is 74 degrees. Then, the application can select a plane arranged opposite to the image to be recognized as a mapping plane, a point arranged opposite to the rotation point in the mapping plane is used as a coordinate origin, a second coordinate system is constructed with the right direction as the X forward direction and the downward direction as the Y forward direction according to the coordinate origin, after the second coordinate system is constructed, the coordinates of four corner points and (0,0), (w, 0), (w, h) and (0,h) are subjected to perspective transformation in a one-to-one correspondence manner, and a new picture with the same size as the original text block in width and height is obtained, wherein the new picture is the text block located in the horizontal direction in the application, and w and h are the width and the height of the text block respectively.
In one embodiment, as shown in fig. 3 and 4, fig. 3 is a text block located in a horizontal direction according to an embodiment of the present application, and fig. 4 is another text block located in a horizontal direction according to an embodiment of the present application; the identifying the text block located in the horizontal direction in S130 may include:
s301: determining whether the text in the horizontally-located text block is inverted; if yes, executing S302; if not, go to step S303.
S302: and performing 180-degree rotation operation on the text block in the horizontal direction, and identifying the rotated text block.
S303: and directly recognizing the text block in the horizontal direction.
In this embodiment, after the text block in the horizontal direction is obtained, the present application may further continue to determine whether the text in the text block is inverted, if so, perform a 180 ° rotation operation on the text block in the horizontal direction, and identify the text block after the rotation, and if not, may directly identify the text block in the horizontal direction.
It can be understood that, by the text recognition method of the present application, the text block in the horizontal direction is obtained after correcting the direction of the text block in the image to be recognized, but the text block may also be as shown in fig. 3 and 4, that is, the text in the text block in the horizontal direction may or may not be inverted, so that the present application may detect the text in the text block in the horizontal direction, confirm whether the text is inverted, and if the text is inverted, perform 180 ° rotation operation on the text block, and recognize the text block after rotation, which may further improve the text recognition efficiency and recognition accuracy.
In one embodiment, the determining whether the text in the text block located in the horizontal direction is inverted in S301 may include: and inputting the text block positioned in the horizontal direction into a pre-configured text recognition model to obtain the inversion condition of the text in the text block positioned in the horizontal direction output by the text recognition model.
In this embodiment, when determining whether the text in the text block located in the horizontal direction is inverted, the text block may be input into a pre-configured text recognition model configured to: the method comprises the steps of taking a training text block positioned in the horizontal direction as a training sample, and taking the inversion condition of a text marked in the training text block as a sample label for training, so that the inversion condition of the text in the text block output by a text recognition model can be obtained by inputting the text block into the trained text recognition model, wherein the inversion condition can comprise inversion and non-inversion, if the text is inverted, the text block positioned in the horizontal direction can be rotated by 180 degrees, the rotated text block can be recognized, and if the text block is not inverted, the text block positioned in the horizontal direction can be directly recognized.
It can be understood that, since the text has a time sequence characteristic, it can be recognized whether the text is inverted or not under the time sequence characteristic. Therefore, the text recognition model in the application can adopt convolutional layer extraction characteristics, then is accessed to a recurrent neural network to build a deep learning crnn model to recognize whether the text in the text block is inverted or not, so that two-classification training is carried out, and finally the trained model is used as an inference model to judge whether the text in the input text picture is inverted or not.
In the following, the text recognition apparatus provided in the embodiment of the present application is described, and the text recognition apparatus described below and the text recognition method described above may be referred to correspondingly.
In an embodiment, as shown in fig. 5, fig. 5 is a schematic structural diagram of a text recognition apparatus provided in an embodiment of the present application; the present application further provides a text recognition apparatus, which may include a text obtaining module 210, a rotation parameter determining module 220, and a text recognition module 230, and specifically includes the following:
the text obtaining module 210 is configured to obtain a text block in a specified area in an image to be recognized and position information of the text block, where the position information of the text block includes coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be recognized.
A rotation parameter determining module 220, configured to determine, according to coordinates of four corner points of the text block, a rotation point and a rotation angle for rotating the text block to a horizontal direction.
And the text recognition module 230 is configured to rotate the text block to the horizontal direction according to the rotation point and the rotation angle, and recognize the text block located in the horizontal direction.
In the above embodiment, before the text in the image to be recognized is recognized, the text block in the designated area in the image to be recognized and the position information of the text block may be obtained, where the position information may include coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be recognized, and then the application may determine, according to the coordinates corresponding to the four corner points of the text block, a rotation point and a rotation angle for rotating the text block to the horizontal direction, for example, the application may determine, according to the coordinates corresponding to the four corner points of the text block, a broadside and a rotation angle of the broadside of the text block, and then take any point on the broadside as a rotation point, and rotate the text block to the horizontal direction according to the rotation point and the rotation angle, thereby recognizing the text block located in the horizontal direction; in the process, the direction of the text block is corrected before the text block in the image to be recognized is recognized, so that the direction of the text in the text block to be recognized is fixed, the text recognition efficiency can be improved, and the accuracy of the text recognition can be improved.
In one embodiment, the rotation parameter determining module 220 may include:
and the width and height determining module is used for determining the width and height of the text block according to the coordinates of the four corner points of the text block, wherein the width of the text block is not less than the height of the text block.
The coordinate sorting module is used for traversing the coordinates of the four corner points of the text block and sorting the coordinates of the four corner points clockwise or anticlockwise according to the width and the height of the text block; among the four corner points after sequencing, the distance from the first corner point to the second corner point is consistent with the width of the text block, and the distance from the second corner point to the third corner point is consistent with the height of the text block.
And the rotation point determining module is used for taking the first corner point of the sequenced four corner points as a rotation point for rotating the text block to the horizontal direction.
In one embodiment, the coordinate sorting module may include:
and the starting point determining module is used for traversing the coordinates of the four corner points of the text block and taking the corner point with the smallest sum of the horizontal coordinate and the vertical coordinate in the four corner points of the text block as the starting point.
And the first sorting module is used for sorting the coordinates of the remaining corner points in the four corner points from the starting point in a clockwise or anticlockwise direction to obtain a sorting result.
And the distance determining module is used for determining whether the distance from the first corner point to the second corner point in the sequencing result is consistent with the width of the text block or not and whether the distance from the second corner point to the third corner point in the sequencing result is consistent with the height of the text block or not according to the width of the text block, the distance from the first corner point to the second corner point in the sequencing result and the distance from the second corner point to the third corner point in the sequencing result.
And a reordering module, configured to reorder, if there is an inconsistency, coordinates of four corner points in the ordering result according to an ordering direction in a last ordering, and return to execute a step of determining whether a distance between a first corner point and a second corner point in the ordering result is consistent with a width of the text block or not and whether a distance between a second corner point and a third corner point in the ordering result is consistent with a height of the text block or not, according to a width and a height of the text block after obtaining a new ordering result, and a distance between a first corner point and a second corner point and a distance between a second corner point and a third corner point in the ordering result are consistent with the height of the text block or not, until the results are consistent.
And the result determining module is used for taking the sorting result as a final sorting result if the results are consistent.
In one embodiment, the rotation parameter determining module 220 may include:
and the broadside determining module is used for determining the broadside of the text block according to the coordinates of the four corner points of the text block.
And the rotation angle determining module is used for calculating an included angle between the wide edge of the text block and the coordinate axis of the first coordinate system, and determining a rotation angle when the text block is rotated to the horizontal direction according to the included angle.
In one embodiment, the text recognition module 230 may include:
and the mapping plane determining module is used for taking a plane which is arranged opposite to the image to be recognized as a mapping plane.
And the coordinate system determining module is used for taking a point which is arranged in the mapping plane and is opposite to the rotating point as a coordinate origin and constructing a second coordinate system according to the coordinate origin.
And the perspective transformation module is used for performing perspective transformation on the text block under the second coordinate system according to the rotation angle by taking the coordinate origin as a coordinate starting point of the text block after the perspective transformation to obtain the text block in the horizontal direction.
In one embodiment, the text recognition module 230 may include:
and the inversion confirmation module is used for determining whether the text in the text block in the horizontal direction is inverted.
And the rotation module is used for performing 180-degree rotation operation on the text block in the horizontal direction if the text block is in the horizontal direction, and identifying the rotated text block.
And the direct identification module is used for directly identifying the text block in the horizontal direction if the text block is not in the horizontal direction.
In one embodiment, the inversion confirmation module may include:
and the model prediction module is used for inputting the text block in the horizontal direction into a pre-configured text recognition model to obtain the inversion condition of the text in the text block in the horizontal direction output by the text recognition model.
The text recognition model is obtained by training with a training text block positioned in the horizontal direction as a training sample and with the inversion condition of the text marked in the training text block as a sample label.
In one embodiment, the present application further provides a storage medium having computer-readable instructions stored therein, which when executed by one or more processors, cause the one or more processors to perform the steps of the text recognition method as described in any one of the above embodiments.
In one embodiment, the present application further provides a computer device comprising: one or more processors, and a memory.
The memory has stored therein computer readable instructions which, when executed by the one or more processors, perform the steps of the text recognition method of any of the above embodiments.
Fig. 6 is a schematic diagram illustrating an internal structure of a computer device according to an embodiment of the present disclosure, and the computer device 300 may be provided as a server, as shown in fig. 6. Referring to fig. 6, computer device 300 includes a processing component 302 that further includes one or more processors, and memory resources, represented by memory 301, for storing instructions, such as application programs, that are executable by processing component 302. The application program stored in memory 301 may include one or more modules that each correspond to a set of instructions. Further, the processing component 302 is configured to execute instructions to perform the text recognition method of any of the embodiments described above.
The computer device 300 may also include a power component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input output (I/O) interface 305. The computer device 300 may operate based on an operating system stored in memory 301, such as Windows Server, mac OS XTM, unix, linux, free BSDTM, or the like.
It will be appreciated by those skilled in the art that the configuration shown in fig. 6 is a block diagram of only a portion of the configuration associated with the present application, and is not intended to limit the computing device to which the present application may be applied, and that a particular computing device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, the embodiments may be combined as needed, and the same and similar parts may be referred to each other.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method of text recognition, the method comprising:
acquiring a text block of a specified area in an image to be recognized and position information of the text block, wherein the position information of the text block comprises coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be recognized;
determining a rotation point and a rotation angle for rotating the text block to the horizontal direction according to the coordinates of the four corner points of the text block;
and rotating the text block to the horizontal direction according to the rotation point and the rotation angle, and identifying the text block in the horizontal direction.
2. The method of claim 1, wherein determining rotation points for rotating the text block to a horizontal direction according to coordinates of four corner points of the text block comprises:
determining the width and the height of the text block according to the coordinates of four corner points of the text block, wherein the width of the text block is not less than the height of the text block;
traversing coordinates of four corner points of the text block, and sequencing the coordinates of the four corner points clockwise or anticlockwise according to the width and the height of the text block; among the four corner points after sequencing, the distance from a first corner point to a second corner point is consistent with the width of the text block, and the distance from the second corner point to a third corner point is consistent with the height of the text block;
and taking the first corner point of the four sorted corner points as a rotation point for rotating the text block to the horizontal direction.
3. The text recognition method of claim 2, wherein traversing coordinates of four corner points of the text block and sorting the coordinates of the four corner points clockwise or counterclockwise according to the width and height of the text block comprises:
traversing coordinates of four edge points of the text block, and taking the edge point with the smallest sum of the abscissa and the ordinate in the four edge points of the text block as a starting point;
from the starting point, sorting the coordinates of the remaining corner points in the four corner points in a clockwise or anticlockwise direction to obtain a sorting result;
determining whether the distance from the first corner point to the second corner point in the sequencing result is consistent with the width of the text block or not and whether the distance from the second corner point to the third corner point in the sequencing result is consistent with the height of the text block or not according to the width of the text block, the distance from the first corner point to the second corner point in the sequencing result and the distance from the second corner point to the third corner point in the sequencing result;
if the two edge points are inconsistent, re-ordering the coordinates of the four edge points in the ordering result according to the ordering direction in the last ordering, returning to execute the step of determining whether the distance from the first edge point to the second edge point in the ordering result is consistent with the width of the text block or not and whether the distance from the second edge point to the third edge point in the ordering result is consistent with the height of the text block or not according to the width of the text block and the distance from the first edge point to the second edge point and the distance from the second edge point to the third edge point in the ordering result are consistent with the height of the text block or not after obtaining a new ordering result;
and if the two are consistent, taking the sorting result as a final sorting result.
4. The method for recognizing text according to claim 1, wherein the determining rotation angles for rotating the text block to a horizontal direction according to coordinates of four corner points of the text block comprises:
determining the broadside of the text block according to the coordinates of the four corner points of the text block;
and calculating an included angle between the wide side of the text block and the coordinate axis of the first coordinate system, and determining a rotation angle when the text block is rotated to the horizontal direction according to the included angle.
5. The text recognition method of claim 1, wherein the rotating the text block to a horizontal direction according to the rotation point and the rotation angle to obtain a text block located in the horizontal direction comprises:
taking a plane which is arranged opposite to the image to be recognized as a mapping plane;
taking a point, which is arranged in the mapping plane and is opposite to the rotation point, as a coordinate origin, and constructing a second coordinate system according to the coordinate origin;
and performing perspective transformation on the text block to the second coordinate system according to the rotation angle by taking the coordinate origin as the coordinate starting point of the text block after the perspective transformation to obtain the text block in the horizontal direction.
6. The method of claim 1, wherein the identifying the text block located in the horizontal direction comprises:
determining whether the text in the horizontally-located text block is inverted;
if so, performing 180-degree rotation operation on the text block in the horizontal direction, and identifying the rotated text block;
if not, directly identifying the text block in the horizontal direction.
7. The method of claim 6, wherein the determining whether the text in the horizontally-oriented text block is inverted comprises:
inputting the text block in the horizontal direction into a pre-configured text recognition model to obtain the inversion condition of the text in the text block in the horizontal direction output by the text recognition model;
the text recognition model is obtained by training with a training text block positioned in the horizontal direction as a training sample and with the inversion condition of the text marked in the training text block as a sample label.
8. A text recognition apparatus, comprising:
the system comprises a text acquisition module, a recognition module and a recognition module, wherein the text acquisition module is used for acquiring a text block of a specified area in an image to be recognized and position information of the text block, and the position information of the text block comprises coordinates of four corner points of the text block in a first coordinate system corresponding to the image to be recognized;
the rotation parameter determining module is used for determining a rotation point and a rotation angle for rotating the text block to the horizontal direction according to the coordinates of the four corner points of the text block;
and the text recognition module is used for rotating the text block to the horizontal direction according to the rotation point and the rotation angle and recognizing the text block in the horizontal direction.
9. A storage medium, characterized by: the storage medium having stored therein computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the text recognition method of any one of claims 1 to 7.
10. A computer device, comprising: one or more processors, and a memory;
the memory has stored therein computer-readable instructions that, when executed by the one or more processors, perform the steps of the text recognition method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210885587.3A CN115147843A (en) | 2022-07-26 | 2022-07-26 | Text recognition method and device, storage medium and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210885587.3A CN115147843A (en) | 2022-07-26 | 2022-07-26 | Text recognition method and device, storage medium and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115147843A true CN115147843A (en) | 2022-10-04 |
Family
ID=83414196
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210885587.3A Pending CN115147843A (en) | 2022-07-26 | 2022-07-26 | Text recognition method and device, storage medium and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115147843A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117315664A (en) * | 2023-09-18 | 2023-12-29 | 山东博昂信息科技有限公司 | Scrap steel bucket number identification method based on image sequence |
-
2022
- 2022-07-26 CN CN202210885587.3A patent/CN115147843A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117315664A (en) * | 2023-09-18 | 2023-12-29 | 山东博昂信息科技有限公司 | Scrap steel bucket number identification method based on image sequence |
| CN117315664B (en) * | 2023-09-18 | 2024-04-02 | 山东博昂信息科技有限公司 | Scrap steel bucket number identification method based on image sequence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110307838B (en) | Robot repositioning method and device, computer-readable storage medium and robot | |
| US6278798B1 (en) | Image object recognition system and method | |
| CN109948590B (en) | Attitude problem detection method and device | |
| JP2885823B2 (en) | Visual recognition device | |
| CN109815770B (en) | Two-dimensional code detection method, device and system | |
| US20100104194A1 (en) | Image processing apparatus, electronic medium, and image processing method | |
| CN108428248B (en) | Vehicle window positioning method, system, equipment and storage medium | |
| US20080084413A1 (en) | Method for generating multi-resolution three-dimensional model | |
| CN111461113B (en) | Large-angle license plate detection method based on deformed plane object detection network | |
| WO2017219992A1 (en) | Indoor navigation method for handheld terminal, handheld terminal, and storage medium | |
| US11176425B2 (en) | Joint detection and description systems and methods | |
| CN112651490B (en) | Training method and device for human face key point detection model and readable storage medium | |
| CN111428717A (en) | Text recognition method and device, electronic equipment and computer readable storage medium | |
| CN112163588A (en) | Intelligent evolution-based heterogeneous image target detection method, storage medium and equipment | |
| CN106778587B (en) | State of flight detection method and device based on unmanned plane image | |
| CN110660089A (en) | Satellite image registration method and device | |
| CN113205041A (en) | Structured information extraction method, device, equipment and storage medium | |
| CN115147843A (en) | Text recognition method and device, storage medium and computer equipment | |
| JP4521568B2 (en) | Corresponding point search method, relative orientation method, three-dimensional image measurement method, corresponding point search device, relative orientation device, three-dimensional image measurement device, corresponding point search program, and computer-readable recording medium recording the corresponding point search program | |
| CN108961164A (en) | Image registration joining method and device and computer equipment based on geometrical invariants | |
| CN111105435A (en) | Marker matching method and device and terminal equipment | |
| CN113673288A (en) | Idle parking space detection method and device, computer equipment and storage medium | |
| CN104616302A (en) | Real-time object identification method | |
| CN116052175A (en) | Text detection method, electronic device, storage medium and computer program product | |
| CN114359222A (en) | Method for detecting arbitrary polygon target, electronic device 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 |