CN111382750A - Method and device for identifying graphic verification code - Google Patents
Method and device for identifying graphic verification code Download PDFInfo
- Publication number
- CN111382750A CN111382750A CN202010148180.3A CN202010148180A CN111382750A CN 111382750 A CN111382750 A CN 111382750A CN 202010148180 A CN202010148180 A CN 202010148180A CN 111382750 A CN111382750 A CN 111382750A
- Authority
- CN
- China
- Prior art keywords
- verification code
- verification
- identification model
- training
- generating
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Character Discrimination (AREA)
Abstract
The invention provides a method and a device for identifying a pattern verification code, wherein the method comprises the following steps: generating a predetermined number of tagged analog validation codes; inputting the generated simulation verification code into a deep learning network for training to obtain a verification code identification model; and carrying out transfer learning on the verification code identification model based on the marked real verification code. In the invention, the artificial marking of the verification code label is reduced by generating the simulation verification code with the label, and the verification code identification model is further migrated and learned, so that the verification code identification model can better identify the verification code, and the effect of improving the identification accuracy of the verification code is achieved on the basis of saving the cost of artificial marking.
Description
Technical Field
The invention relates to the field of identifying codes, in particular to a method and a device for identifying a graphic identifying code.
Background
With the development of deep learning, the technology of recognizing characters in a picture with characters is more and more mature. If only one character exists in the picture, ten thousand characters written in different fonts can be recognized with higher accuracy by training the character in the picture through a deep learning method.
If a picture contains a plurality of characters, most of the time, the picture needs to be divided into pictures with single characters, of course, the process of dividing the picture with the plurality of characters into the plurality of pictures only containing single characters has errors, and the more the noise of the verification code is, the more complicated the noise is, the poorer the dividing effect is. Although it is easier to recognize the content in a picture with a single character than to recognize the whole picture, the quality of the segmentation becomes a key factor affecting the recognition rate. Another method for identifying all characters in a picture containing multiple characters is to directly identify the whole picture, and the method has no segmentation error, and the size of data volume and an identification algorithm become key factors influencing the identification rate.
Disclosure of Invention
The embodiment of the invention provides a method and a device for identifying a graph verification code, which at least solve the problem of low accuracy of a graph verification code identification mode based on a deep learning network in the related art.
According to an embodiment of the present invention, there is provided a pattern verification code recognition method including: generating a predetermined number of tagged analog validation codes; inputting the generated simulation verification code into a deep learning network for training to obtain a verification code identification model; and carrying out transfer learning on the verification code identification model based on the marked real verification code.
Optionally, the generating a predetermined number of tagged analog validation codes may include: a true validation code is collected and a generative countermeasure network is trained based on the true validation code to generate a predetermined number of simulated validation codes that are similar to the true validation code.
Optionally, the inputting the generated simulated verification code into a deep learning network for training to obtain a verification code recognition model may include: the generated simulation verification codes are combined into a simulation verification code training data set; and training a deep learning network by taking the simulated verification code training data set as input to obtain the verification code identification model.
Optionally, after the migration learning of the verification code identification model based on the labeled real verification code, the method includes: and identifying the graphic verification code based on the verification code identification model.
According to another embodiment of the present invention, there is provided a graphic verification code recognition apparatus including: the generating module is used for generating a preset number of labeled simulation verification codes; the acquisition module is used for inputting the generated simulated verification code into a deep learning network for training to obtain a verification code identification model; and the transfer learning module is used for carrying out transfer learning on the verification code identification model based on the marked real verification code.
Optionally, the generating module may include: a collecting unit for collecting the true verification code; and the generating unit is used for training the generating countermeasure network based on the real verification code so as to generate a preset number of simulation verification codes similar to the real verification code.
Optionally, the obtaining module may include: the composition unit is used for composing the generated analog verification code into an analog verification code training data set; and the acquisition unit is used for training a deep learning network by taking the simulated verification code training data set as input to obtain the verification code identification model.
Optionally, the apparatus may further include: and the identification module is used for identifying the graphic verification code based on the verification code identification model.
According to a further embodiment of the present invention, there is also provided a storage medium having a computer program stored therein, wherein the computer program is arranged to perform the steps of any of the above method embodiments when executed.
According to yet another embodiment of the present invention, there is also provided an electronic device, including a memory in which a computer program is stored and a processor configured to execute the computer program to perform the steps in any of the above method embodiments.
In the embodiment of the invention, the number of the verification code labels marked manually is reduced by generating the simulation verification code with the label, and the verification code identification model is further migrated and learned, so that the verification code identification model can better identify the verification code, and the effect of improving the verification code identification accuracy is achieved on the basis of saving the cost of manual marking.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:
FIG. 1 is a flow chart of a graphical passcode identification method according to an embodiment of the invention;
FIG. 2 is a block diagram of a graphic verification code recognition apparatus according to an embodiment of the present invention;
FIG. 3 is a block diagram of a graphical authentication code identification apparatus according to an alternative embodiment of the present invention;
FIG. 4 is a flow diagram of a method for implementing a graphical verification code identification technique according to an embodiment of the present invention;
FIG. 5 is a flow chart of a combined alpha and numeric validation code acquisition, labeling, training recognition method according to an alternate embodiment of the present invention;
FIG. 6 is a diagram of the recognition effect of the verification code for the southern China airline website according to an alternative embodiment of the present invention;
fig. 7 is a diagram illustrating the recognition effect of the verification code of the website of the china construction bank according to an alternative embodiment of the present invention.
Detailed Description
The invention will be described in detail hereinafter with reference to the accompanying drawings in conjunction with embodiments. It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
Example 1
In this embodiment, a method for identifying a pattern verification code is provided, and fig. 1 is a flowchart of the method for identifying a pattern verification code according to an embodiment of the present invention, as shown in fig. 1, the flowchart includes the following steps:
step S101, generating a preset number of labeled simulation verification codes;
step S102, inputting the generated simulation verification code into a deep learning network for training to obtain a verification code identification model;
and step S103, performing transfer learning on the verification code identification model based on the marked real verification code.
In this embodiment, step S101 may include: a true validation code is collected and a generative countermeasure network is trained based on the true validation code to generate a predetermined number of simulated validation codes that are similar to the true validation code.
In this embodiment, step S102 may include: the generated simulation verification codes are combined into a simulation verification code training data set; and training a deep learning network by taking the simulated verification code training data set as input to obtain the verification code identification model.
After step S103 of this embodiment, the method may further include: and identifying the graphic verification code based on the verification code identification model.
Through the steps, the number of the manually marked verification code labels is reduced by generating the simulated verification code with the label, and the verification code identification model is subjected to further transfer learning, so that the verification code identification model can better identify the verification code, and the effect of improving the identification accuracy of the verification code is achieved on the basis of saving the cost of manual marking.
Through the above description of the embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but the former is a better implementation mode in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, or a network device) to execute the method according to the embodiments of the present invention.
Example 2
In this embodiment, a device for identifying a pattern verification code is further provided, where the device is used to implement the foregoing embodiments and preferred embodiments, and the description already made is omitted for brevity. As used below, the terms "module" and "unit" may implement a combination of software and/or hardware of predetermined functions. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 2 is a block diagram of a configuration of a pattern verification code recognition apparatus according to an embodiment of the present invention, and as shown in fig. 2, the apparatus includes a generation module 10, an acquisition module 20, and a migration learning module 30.
The generation module 10 is used for generating a predetermined number of labeled analog verification codes.
And the obtaining module 20 is configured to input the generated simulated verification code into a deep learning network to train to obtain a verification code identification model.
And the transfer learning module 30 is configured to perform transfer learning on the verification code identification model based on the labeled real verification code.
Fig. 3 is a block diagram showing a configuration of a pattern authentication code recognition apparatus according to an alternative embodiment of the present invention, which may further include an identification module 40 in addition to all the modules shown in fig. 2, as shown in fig. 3. And the identification module 40 is used for identifying the graphic verification code based on the verification code identification model.
In this embodiment, the generating module 10 may further include a collecting unit 11 and a generating unit 12.
And the collecting unit 11 is used for collecting the true verification code.
A generating unit 12, configured to train a generative countermeasure network based on the authentic verification code to generate a predetermined number of simulated verification codes similar to the authentic verification code.
In this embodiment, the obtaining module 20 may further include a composition unit 21 and an obtaining unit 22.
A forming unit 21, configured to form the generated verification code into a training data set of the verification code.
And the obtaining unit 22 is configured to train the deep learning network by using the simulated verification code training data set as an input to obtain the verification code identification model.
It should be noted that, the above modules may be implemented by software or hardware, and for the latter, the following may be implemented, but not limited to: the modules are all positioned in the same processor; alternatively, the modules are respectively located in different processors in any combination.
Example 3
In order to facilitate understanding of the technical solutions provided by the present invention, the following detailed description will be made with reference to embodiments of specific scenarios.
For the verification code generated by randomly combining letters and numbers, noise is artificially and deliberately increased in the generation process of the verification code, a part where two connected characters in the verification code are overlapped may exist, and the characters in the verification code may also be distorted, so that the error rate is high if the whole verification code is divided into pictures only containing single characters. The embodiment introduces a method for directly identifying all characters in a verification code without segmenting a verification code picture, and compares the verification code identification method through an open-source general OCR, wherein the open-source general OCR is a method for identifying characters in the verification code by segmenting before identifying. The accuracy rate of the OCR identifying verification code based on the open source is only 66%, and the identifying verification code identified by the technology of segmenting firstly and then identifying has lower accuracy rate.
The method based on deep learning is used for identifying all characters in the whole verification code, so that a large number of verification codes need to be collected, and the characters in the verification codes are labeled to be used as a training set with labels, so that supervised learning is performed. Noise is intentionally added by human in the process of generating the verification code, so that the quantity of the verification code data required to be labeled by training the deep learning network cannot be estimated. It is not enough to mark more or less complicated verification codes. This method requires a large amount of time and labor cost to implement.
The process of identifying the verification code introduced in this embodiment includes four major parts, namely, collecting the verification code, labeling the verification code, generating a large number of similar verification codes, and training the identification verification code. The method described in this embodiment aims to achieve a higher recognition accuracy by labeling fewer verification codes.
The present embodiment aims to achieve a higher recognition rate by labeling fewer authentication codes. Three technical implementation methods are mainly adopted. Fig. 4 is a flowchart of a method for implementing a pattern verification code recognition technology according to an embodiment of the present invention, as shown in fig. 4, the method includes the following steps:
step S401, generating a large number of verification codes with labels and similar to the real verification codes;
step S402, training the generated large amount of labeled verification codes and generating a model;
step S403, on the basis of the model, migrating and learning the labeled real verification code with the label to obtain a higher recognition rate.
Taking the verification code of a certain airline website as an example, based on the three main technologies shown in fig. 4, only 500 verification codes need to be marked and sent into the model for transfer learning, the final accuracy rate is 98%, and only 0.04s is needed for identifying one verification code. Compared with the method of firstly segmenting and then identifying, the accuracy rate is improved from about 66% to 98%, compared with the method of manually labeling a large number of pictures and then identifying the whole verification code, the cost of manual labeling is greatly saved, and the time required for labeling 500 verification codes containing 4 characters by a single person is about 2 hours.
Fig. 5 is a flowchart of a method for collecting, labeling, training and recognizing a verification code by combining letters and numbers according to an alternative embodiment of the present invention, and as shown in fig. 5, the flowchart includes the following steps:
and step S501, collecting the verification code.
The embodiment collects the verification codes of two websites, namely a website of a certain airline company and a website of a certain bank. The verification code of the website of the airline company is a 4-character verification code, that of the website of the bank is a 5-character verification code, and the verification code of the website of the bank is more distorted.
And step S502, marking the verification code.
The time for manually marking the 4-character verification code of 500 airline websites is about 2 hours. The time for manually marking the 5-character verification code of 500 bank websites is about 2 and a half hours.
In step S503, a plurality of similar verification codes are generated.
In this embodiment, a large number of verification codes similar to the true verification code are generated based on a Generic Adaptive Network (GAN) algorithm, so as to reduce the number of verification codes that need to be manually labeled. A generator in the trained GAN model can continuously generate a large number of similar verification codes, and the requirements of training a deep learning model are met.
Step S504, training and identifying the verification code.
And sending a large number of generated similar verification codes into a deep learning network for training to obtain a verification code identification model. On the basis of the model, transfer learning is carried out based on the marked real verification code so as to improve the recognition rate.
In the embodiment of the invention, the identification rate of the verification code is improved as much as possible on the basis of saving the cost of manual marking as much as possible. By generating a large number of similar identifying code pictures with labels, the deep learning is ensured, and a deep learning network with high accuracy can be trained. In addition, migration learning is carried out on the basis of the true identifying code picture marked with the label, so that the deep learning network can better identify the identifying code.
Through the steps, the cost of manually marking the verification code is greatly saved on the basis of ensuring the identification accuracy, and the identification accuracy of the verification code can reach 98% for an airline website and a bank website.
Compared with the traditional method for recognizing characters in the verification code by using OCR, the method has the advantages that the accuracy is improved, and the accuracy is improved from 66% to 98%.
Compared with deep learning which needs to label a large number of verification code pictures, the method greatly saves the cost of manually labeling the verification code labels, only 500 verification codes need to be labeled, and manual labeling can be completed within about 2 hours.
The identification rate of the verification codes of the airline website and the bank website reaches 98 percent. The experimental results are as follows.
And (4) randomly and newly acquiring and marking the verification codes of 500 airline websites, and identifying the verification codes based on the trained model. The recognition effect of the 500 verification code is shown in fig. 6, and fig. 6 is a diagram of the recognition effect of the verification code of the airline website according to an alternative embodiment of the present invention.
As shown in fig. 6, the erroneous pictures are analyzed, and most of the characters (B,8), (B,3), (K, X), (S,5) are recognized as errors, and some of the pictures may not be recognized correctly by naked eyes.
And (5) newly acquiring and labeling 500 verification codes of the bank website at random, and identifying the verification codes based on the trained model. The identification effect of the 500 authentication code is shown in fig. 7, and fig. 7 is a diagram illustrating the identification effect of the authentication code of the bank website according to an alternative embodiment of the present invention.
As shown in fig. 7, the erroneous pictures are analyzed, and most of the characters (m, n), (t, f), (3, p) are recognized as errors, and some of the pictures cannot be recognized correctly by naked eyes.
The test results show that the method introduced by the embodiment can effectively reduce the manual labeling cost and improve the identification accuracy of the verification code.
Example 4
Embodiments of the present invention also provide a storage medium having a computer program stored therein, wherein the computer program is arranged to perform the steps of any of the above method embodiments when executed.
Alternatively, in the present embodiment, the storage medium may be configured to store a computer program for executing the steps of:
s1, generating a preset number of labeled simulation verification codes;
s2, inputting the generated simulated verification code into a deep learning network for training to obtain a verification code identification model;
and S3, performing transfer learning on the verification code identification model based on the marked real verification code.
Optionally, in this embodiment, the storage medium may include, but is not limited to: various media capable of storing computer programs, such as a usb disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic disk, or an optical disk.
Example 5
Embodiments of the present invention also provide an electronic device comprising a memory having a computer program stored therein and a processor arranged to run the computer program to perform the steps of any of the above method embodiments.
Optionally, the electronic apparatus may further include a transmission device and an input/output device, wherein the transmission device is connected to the processor, and the input/output device is connected to the processor.
Optionally, in this embodiment, the processor may be configured to execute the following steps by a computer program:
s1, generating a preset number of labeled simulation verification codes;
s2, inputting the generated simulated verification code into a deep learning network for training to obtain a verification code identification model;
and S3, performing transfer learning on the verification code identification model based on the marked real verification code.
Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments and optional implementation manners, and this embodiment is not described herein again.
It will be apparent to those skilled in the art that the modules or steps of the present invention described above may be implemented by a general purpose computing device, they may be centralized on a single computing device or distributed across a network of multiple computing devices, and alternatively, they may be implemented by program code executable by a computing device, such that they may be stored in a storage device and executed by a computing device, and in some cases, the steps shown or described may be performed in an order different than that described herein, or they may be separately fabricated into individual integrated circuit modules, or multiple ones of them may be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A method for identifying a pattern verification code, comprising:
generating a predetermined number of tagged analog validation codes;
inputting the generated simulation verification code into a deep learning network for training to obtain a verification code identification model;
and carrying out transfer learning on the verification code identification model based on the marked real verification code.
2. The method of claim 1, wherein generating a predetermined number of tagged analog validation codes comprises:
a true validation code is collected and a generative countermeasure network is trained based on the true validation code to generate a predetermined number of simulated validation codes that are similar to the true validation code.
3. The method of claim 1, wherein inputting the generated simulated verification code into a deep learning network for training to obtain a verification code recognition model comprises:
the generated simulation verification codes are combined into a simulation verification code training data set;
and training a deep learning network by taking the simulated verification code training data set as input to obtain the verification code identification model.
4. The method of claim 1, wherein after the step of performing the migration learning on the verification code recognition model based on the labeled real verification code, the method further comprises:
and identifying the graphic verification code based on the verification code identification model.
5. An apparatus for identifying a pattern verification code, comprising:
the generating module is used for generating a preset number of labeled simulation verification codes;
the acquisition module is used for inputting the generated simulated verification code into a deep learning network for training to obtain a verification code identification model;
and the transfer learning module is used for carrying out transfer learning on the verification code identification model based on the marked real verification code.
6. The apparatus of claim 5, wherein the generating module comprises:
a collecting unit for collecting the true verification code;
and the generating unit is used for training the generating countermeasure network based on the real verification code so as to generate a preset number of simulation verification codes similar to the real verification code.
7. The apparatus of claim 5, wherein the obtaining module comprises:
the composition unit is used for composing the generated analog verification code into an analog verification code training data set;
and the acquisition unit is used for training a deep learning network by taking the simulated verification code training data set as input to obtain the verification code identification model.
8. The apparatus of claim 5, further comprising:
and the identification module is used for identifying the graphic verification code based on the verification code identification model.
9. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is arranged to perform the method of any of claims 1 to 4 when executed.
10. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, and wherein the processor is arranged to execute the computer program to perform the method of any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010148180.3A CN111382750A (en) | 2020-03-05 | 2020-03-05 | Method and device for identifying graphic verification code |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010148180.3A CN111382750A (en) | 2020-03-05 | 2020-03-05 | Method and device for identifying graphic verification code |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111382750A true CN111382750A (en) | 2020-07-07 |
Family
ID=71217059
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010148180.3A Pending CN111382750A (en) | 2020-03-05 | 2020-03-05 | Method and device for identifying graphic verification code |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111382750A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111935106A (en) * | 2020-07-23 | 2020-11-13 | 湖北工业大学 | Method for collecting labeled data set by using improved verification code |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109523538A (en) * | 2018-11-21 | 2019-03-26 | 上海七牛信息技术有限公司 | A kind of people counting method and system based on generation confrontation neural network |
| US20190130232A1 (en) * | 2017-10-30 | 2019-05-02 | Monotype Imaging Inc. | Font identification from imagery |
| CN110009057A (en) * | 2019-04-16 | 2019-07-12 | 四川大学 | A kind of graphical verification code recognition methods based on deep learning |
| WO2019136960A1 (en) * | 2018-01-12 | 2019-07-18 | 深圳壹账通智能科技有限公司 | Method and device for crawling website data, storage medium and server |
| CN110033445A (en) * | 2019-04-10 | 2019-07-19 | 司法鉴定科学研究院 | Medicolegal examination automatic identification system and recognition methods based on deep learning |
| GB201911724D0 (en) * | 2019-08-15 | 2019-10-02 | Vision Semantics Ltd | Text based image search |
-
2020
- 2020-03-05 CN CN202010148180.3A patent/CN111382750A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190130232A1 (en) * | 2017-10-30 | 2019-05-02 | Monotype Imaging Inc. | Font identification from imagery |
| WO2019136960A1 (en) * | 2018-01-12 | 2019-07-18 | 深圳壹账通智能科技有限公司 | Method and device for crawling website data, storage medium and server |
| CN109523538A (en) * | 2018-11-21 | 2019-03-26 | 上海七牛信息技术有限公司 | A kind of people counting method and system based on generation confrontation neural network |
| CN110033445A (en) * | 2019-04-10 | 2019-07-19 | 司法鉴定科学研究院 | Medicolegal examination automatic identification system and recognition methods based on deep learning |
| CN110009057A (en) * | 2019-04-16 | 2019-07-12 | 四川大学 | A kind of graphical verification code recognition methods based on deep learning |
| GB201911724D0 (en) * | 2019-08-15 | 2019-10-02 | Vision Semantics Ltd | Text based image search |
Non-Patent Citations (3)
| Title |
|---|
| KWON, HYUN等: "CAPTCHA Image Generation: Two-Step Style-Transfer Learning in Deep Neural Networks", 《SENSORS》 * |
| WANG, PING等: "Simple and Easy: Transfer Learning-Based Attacks to Text CAPTCHA", 《IEEE ACCESS》 * |
| YE GUIXIN等: "Yet Another Text Captcha Solver: A Generative Adversarial Network Based Approach", 《PROCEEDINGS OF THE 2018 ACM SIGSAC CONFERENCE ON COMPUTER AND COMMUNICATIONS SECURITY (CCS\'18)》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111935106A (en) * | 2020-07-23 | 2020-11-13 | 湖北工业大学 | Method for collecting labeled data set by using improved verification code |
| CN111935106B (en) * | 2020-07-23 | 2022-05-13 | 湖北工业大学 | Method for collecting labeled data set by using improved verification code |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110704633B (en) | Named entity recognition method, named entity recognition device, named entity recognition computer equipment and named entity recognition storage medium | |
| CN107844481B (en) | Text recognition error detection method and device | |
| CN111694937A (en) | Interviewing method and device based on artificial intelligence, computer equipment and storage medium | |
| CN112541443B (en) | Invoice information extraction method, invoice information extraction device, computer equipment and storage medium | |
| CN106067019A (en) | The method and device of Text region is carried out for image | |
| WO2021232670A1 (en) | Pcb component identification method and device | |
| CN112988963A (en) | User intention prediction method, device, equipment and medium based on multi-process node | |
| CN112418304B (en) | OCR (optical character recognition) model training method, system and device | |
| CN113656547A (en) | Text matching method, device, equipment and storage medium | |
| CN112951233A (en) | Voice question and answer method and device, electronic equipment and readable storage medium | |
| CN114610855B (en) | Dialogue reply generation method and device, electronic equipment and storage medium | |
| CN114881698A (en) | Advertisement compliance auditing method and device, electronic equipment and storage medium | |
| CN115471775A (en) | Information verification method, device and equipment based on screen recording video and storage medium | |
| CN114639152A (en) | Multi-modal voice interaction method, device, equipment and medium based on face recognition | |
| CN110490056A (en) | The method and apparatus that image comprising formula is handled | |
| CN113627160A (en) | Text error correction method and device, electronic equipment and storage medium | |
| CN112560855B (en) | Image information extraction method and device, electronic equipment and storage medium | |
| CN113705749A (en) | Two-dimensional code identification method, device and equipment based on deep learning and storage medium | |
| CN111382750A (en) | Method and device for identifying graphic verification code | |
| CN112613367A (en) | Bill information text box acquisition method, system, equipment and storage medium | |
| CN107071553A (en) | Method, device and computer readable storage medium for modifying video and voice | |
| CN116704528A (en) | Bill identification verification method, device, computer equipment and storage medium | |
| CN113111869B (en) | Method and system for extracting text picture and description thereof | |
| CN113657279A (en) | Bill image layout analysis method and device | |
| CN114049686A (en) | Signature recognition model training method and device and electronic equipment |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230313 Address after: 410000 Room 502, 504, Jiayuan Office Building, No. 369, Yuelu Avenue, Xianjiahu Street, Yuelu District, Changsha City, Hunan Province - 913 Applicant after: Hunan Huawei Jin'an Enterprise Management Co.,Ltd. Address before: 100080 370m south of Huandao, Yanfu Road, Yancun Town, Fangshan District, Beijing Applicant before: BEIJING WANGZHONG GONGCHUANG TECHNOLOGY CO.,LTD. |
|
| TA01 | Transfer of patent application right |