CN110852359A - Family tree identification method and system based on deep learning - Google Patents
Family tree identification method and system based on deep learning Download PDFInfo
- Publication number
- CN110852359A CN110852359A CN201911035972.3A CN201911035972A CN110852359A CN 110852359 A CN110852359 A CN 110852359A CN 201911035972 A CN201911035972 A CN 201911035972A CN 110852359 A CN110852359 A CN 110852359A
- Authority
- CN
- China
- Prior art keywords
- family tree
- deep learning
- word
- module
- tree image
- 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.)
- Granted
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- 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
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
-
- 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
- G06V30/28—Character recognition specially adapted to the type of the alphabet, e.g. Latin alphabet
- G06V30/287—Character recognition specially adapted to the type of the alphabet, e.g. Latin alphabet of Kanji, Hiragana or Katakana characters
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Animal Behavior & Ethology (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Character Discrimination (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a family tree recognition method and system based on deep learning, which are oriented to the field of family tree digitization and are based on a deep learning method, a family tree data rapid digitization scheme with two deep learning tasks of target position detection and target classification separated is designed, firstly, the positions and the contents of Chinese characters in the family tree are judged by respectively training two convolutional neural networks of target position detection and target classification, and finally, the relationship of characters in the family tree is analyzed through a regular expression to draw the digitized family tree. The family tree recognition scheme based on deep learning not only avoids a large amount of manpower recognition, but also ensures the accuracy of the data digitization result.
Description
Technical Field
The invention relates to a family tree identification method and system based on deep learning.
Background
"national history, local will, family will have a score", as one of three systematic history literature records parallel to the positive history and local will, the present family score has a degree of digitization far behind the former two. The digitization of the positive history has been completed early, the local will is ongoing, and is now completed and enters one third of the total amount in the field of commercialization. The digitalization work of the family tree is not started yet.
The reason is that the digitalization of the family tree has several difficulties:
1. the number of the existing spyware in China is extremely large and is continuously expanded, and manpower and material resources are consumed for extracting a large amount of information in genealogy data by depending on manpower.
2. The family tree often has a large number of rarely-used words, the family tree itself lacks a large number of labeled data sets, most of the existing deep learning data sets do not contain the rarely-used words, and therefore, the existing OCR tools have a large number of mistakes in recognition.
3. The description of the family tree is structural, and the relevance and meaning of the contents of each plate are difficult to identify.
Disclosure of Invention
The invention aims to provide a family tree identification method and system based on deep learning.
In order to solve the above problems, the present invention provides a family tree recognition method based on deep learning, which includes:
acquiring a family tree image;
obtaining the position of each Chinese character in the family tree image through a target position detection network based on deep learning;
and obtaining the content of the Chinese characters in the family tree image through a rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image.
Further, in the above method, before obtaining the position of each chinese character in the family tree image through a target position detection network based on deep learning, the method further includes:
training the target position detection network through a document training set with marked character positions;
before the contents of the Chinese characters in the family tree image are obtained through a rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image, the method further comprises the following steps:
the rare word classification network is trained by generating a training set from the rare word plus culture and noise.
Further, in the above method, training the target location detection network by using a training set of documents with labeled text locations includes:
and training a convolution neural network as the target position detection network through a set of character data sets with position information.
Further, in the above method, training the rarely-used word classification network by generating a training set by adding a character and noise to the rarely-used word comprises:
coding each word in the complex word dictionary and the uncommon word dictionary, making a picture corresponding to each corresponding word by using fonts of different styles based on the coding result, and adding noise into the picture to enhance the extended picture so as to obtain a training set;
and training the training set by adopting a classification neural network to obtain the rarely-used word classification network.
Further, in the method, after obtaining the content of the chinese character in the family tree image through a rare character classification network based on deep learning and the position of each chinese character in the family tree image, the method further includes:
extracting relation words between name information and characters in the family tree image through a regular expression based on the position of each Chinese character and the content of the Chinese character in the family tree image;
and drawing a corresponding family tree diagram based on the relation words between the name information and the characters.
According to another aspect of the present invention, there is also provided a family tree recognition system based on deep learning, including:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring a family tree image;
the second module is used for obtaining the position of each Chinese character in the family tree image through a target position detection network based on deep learning;
and the third module is used for obtaining the content of the Chinese characters in the family tree image through a rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image.
Further, in the system, the second module is further configured to train the target location detection network through a document training set labeled with a text location;
the third module is also used for generating a training set through the uncommon word plus the culture and the noise to train the uncommon word classification network.
Further, in the above system, the second module is configured to train a convolutional neural network as the target position detection network through a set of text data sets with position information.
Further, in the above system, the third module is configured to encode each word in the complex word dictionary and the uncommon word dictionary, and make a picture corresponding to each corresponding word based on the result of the encoding and using fonts of different styles, and add noise to the picture to enhance the extended picture, so as to obtain a training set; and training the training set by adopting a classification neural network to obtain the rarely-used word classification network.
Further, in the system, a fourth module is further included, configured to extract, based on a position of each chinese character and content of the chinese character in the family tree image, a relationship term between name information and a character in the family tree image through a regular expression; and drawing a corresponding family tree diagram based on the relation words between the name information and the characters.
Compared with the prior art, the genealogical file contains traditional words and uncommon words on one hand and lacks a data set, and on the other hand, the association and meaning of the contents of each plate in the genealogy are difficult to be identified by a machine. Therefore, the family tree recognition method carries out family tree recognition in a mode of separating two deep learning tasks of target position detection and target classification. The basic flow is shown in FIG. 1. The method comprises the steps of training a convolutional neural network through deep learning to determine the position of each Chinese character in a family tree, training another convolutional neural network specially aiming at the identification of rare and rare Chinese characters to determine the content of each Chinese character, and finally extracting through a regular expression according to the position and the content of each Chinese character to finally obtain the family tree relationship of the reaction in the family tree.
The invention relates to the field of family tree digitization, and discloses a family tree data rapid digitization scheme with two deep learning tasks of target position detection and target classification separated based on a deep learning method. The family tree recognition scheme based on deep learning not only avoids a large amount of manpower recognition, but also ensures the accuracy of the data digitization result.
Drawings
FIG. 1 is a schematic diagram of a family tree recognition method and system based on deep learning according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating the position detection of a target Chinese character according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating the division of the pedigree Chinese character content structure according to an embodiment of the present invention.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
As shown in fig. 1, the present invention provides a family tree recognition method based on deep learning, which includes:
step S1, acquiring a family tree image;
step S2, obtaining the position of each Chinese character in the family tree image through a target position detection network based on deep learning;
and step S3, obtaining the content of the Chinese characters in the family tree image through a rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image.
The invention designs a rapid family tree data digitization scheme with two deep learning tasks of target position detection and target classification based on a deep convolutional neural network, which is oriented to the family tree digitization field, so that the family tree identification scheme based on deep learning not only reduces and avoids a large amount of manpower identification, but also ensures the accuracy of the data digitization result.
As shown in fig. 2, in an embodiment of the family tree recognition method based on deep learning of the present invention, step S2, before obtaining the position of each chinese character in the family tree image through the target position detection network based on deep learning, further includes:
training the target location detection network through a training set of documents labeled with literal locations
In an embodiment of the family tree recognition method based on deep learning, training the target position detection network through a document training set with marked character positions includes:
and training a convolution neural network as the target position detection network through a set of character data sets with position information.
In this case, the shapes of the Chinese characters are mostly unified, and the positions and the sizes of the Chinese characters can be clearly judged even when the Chinese characters are unknown by considering the thinking mode of the daily human brain because the Chinese characters have relatively fixed characteristics. Therefore, without the pedigree data set, the invention trains a convolutional neural network by using a general set of text data sets with position information to obtain the position information of each word.
The pedigree recognition task does not have the requirement on the real-time performance of the neural network, so the extremely high accuracy is pursued by the invention. The invention aims to provide a novel deep learning scheme, so that the requirements of high recognition rate and high efficiency are met as far as possible under the condition that the number of labeled data sets of the family tree is small at present. The invention adopts any network structure which can be used for training at present with higher recognition rate and relative maturity. The trained network can detect different scales of Chinese characters and their positions, as shown in FIG. 2.
The invention relates to the field of family tree digitization, and discloses a family tree data rapid digitization scheme with two deep learning tasks of target position detection and target classification separated based on a deep convolutional neural network.
The invention provides a brand-new scheme to solve the challenges in the family tree recognition, and adopts a mode of separating two deep learning tasks of target position detection and target classification, wherein the two steps are respectively carried out, so that the training by using a limited labeled family tree data set is avoided, and the recognition accuracy of the final network can be ensured to be far beyond that of mainstream OCR software.
In an embodiment of the family tree recognition method based on deep learning, in step S3, before obtaining the content of the chinese characters in the family tree image through the uncommon character classification network based on deep learning and the position of each chinese character in the family tree image, the method further includes:
the rare word classification network is trained by generating a training set from the rare word plus culture and noise.
In an embodiment of the family tree recognition method based on deep learning, a training set is generated by adding rare words and culture and noise to train the rare word classification network, and the method comprises the following steps:
coding each word in the complex word dictionary and the uncommon word dictionary, making a picture corresponding to each corresponding word by using fonts of different styles based on the coding result, and adding noise into the picture to enhance the extended picture so as to obtain a training set;
and training the training set by adopting a classification neural network to obtain the rarely-used word classification network.
After the position of each character is marked by the convolutional neural network, all that needs to be done is to identify the Chinese character in each frame. Although the classification task performed by using the convolutional neural network is more mature than the target detection task, a good database does not exist for complex words and uncommon words in the family tree, and different texts bring great difference to training results. There are differences in different pedigree printing styles. The invention makes data set by itself, each word in the complex word dictionary and the rare word dictionary is coded, the software makes the corresponding picture of each word, and uses the fonts with different styles to make the pictures, and adds various noises to enhance the expansion data set. After the data set is manufactured, the classification neural network which is mainstream at present is adopted for training, and a good effect is achieved.
In an embodiment of the family tree recognition method based on deep learning, in step S3, after obtaining the content of the chinese characters in the family tree image through the uncommon word classification network based on deep learning and the position of each chinese character in the family tree image, the method further includes:
extracting relation words between name information and characters in the family tree image through a regular expression based on the position of each Chinese character and the content of the Chinese character in the family tree image;
and drawing a corresponding family tree diagram based on the relation words between the name information and the characters.
After the position information and the content of the Chinese characters are obtained through the deep learning of the first two steps, the content of each region can be divided through a customized algorithm according to different types of family tree as shown in fig. 3. And extracting related words between name information and figures in the family tree through a regular expression in the family tree, and finally drawing a family tree graph to finish the digitalization of the family tree.
According to another aspect of the present invention, there is also provided a family tree recognition system based on deep learning, including:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring a family tree image;
the second module is used for obtaining the position of each Chinese character in the family tree image through a target position detection network based on deep learning;
and the third module is used for obtaining the content of the Chinese characters in the family tree image through a rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image.
Further, in the system, the second module is further configured to train the target location detection network through a document training set labeled with a text location;
the third module is also used for generating a training set through the uncommon word plus the culture and the noise to train the uncommon word classification network.
Further, in the above system, the second module is configured to train a convolutional neural network as the target position detection network through a set of text data sets with position information.
Further, in the above system, the third module is configured to encode each word in the complex word dictionary and the uncommon word dictionary, and make a picture corresponding to each corresponding word based on the result of the encoding and using fonts of different styles, and add noise to the picture to enhance the extended picture, so as to obtain a training set; and training the training set by adopting a classification neural network to obtain the rarely-used word classification network.
Further, in the system, a fourth module is further included, configured to extract, based on a position of each chinese character and content of the chinese character in the family tree image, a relationship term between name information and a character in the family tree image through a regular expression; and drawing a corresponding family tree diagram based on the relation words between the name information and the characters.
In summary, genealogical files contain complicated words and rare words and lack data sets on one hand, and on the other hand, the association and meaning of the contents of each plate in the genealogy are difficult to be identified by machines. Therefore, the family tree recognition method carries out family tree recognition in a mode of separating two deep learning tasks of target position detection and target classification. The basic flow is shown in FIG. 1. The method comprises the steps of training a convolutional neural network through deep learning to determine the position of each Chinese character in a family tree, training another convolutional neural network specially aiming at the identification of rare and rare Chinese characters to determine the content of each Chinese character, and finally extracting through a regular expression according to the position and the content of each Chinese character to finally obtain the family tree relationship of the reaction in the family tree.
The invention relates to the field of family tree digitization, and discloses a family tree data rapid digitization scheme with two deep learning tasks of target position detection and target classification separated based on a deep learning method. The family tree recognition scheme based on deep learning not only avoids a large amount of manpower recognition, but also ensures the accuracy of the data digitization result.
For details of each system embodiment of the present invention, reference may be made to corresponding parts of each method embodiment, and details are not described herein again.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (10)
1. A family tree recognition method based on deep learning is characterized by comprising the following steps:
acquiring a family tree image;
obtaining the position of each Chinese character in the family tree image through a target position detection network based on deep learning;
and obtaining the content of the Chinese characters in the family tree image through a rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image.
2. The method for identifying family tree based on deep learning of claim 1, wherein before obtaining the position of each Chinese character in the family tree image through the target position detection network based on deep learning, the method further comprises:
training the target position detection network through a document training set with marked character positions;
before the contents of the Chinese characters in the family tree image are obtained through a rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image, the method further comprises the following steps:
the rare word classification network is trained by generating a training set from the rare word plus culture and noise.
3. The method for family tree recognition based on deep learning of claim 2, wherein training the target position detection network through a training set of documents with labeled literal positions comprises:
and training a convolution neural network as the target position detection network through a set of character data sets with position information.
4. The method for deep learning based pedigree recognition of claim 2, wherein training the rarely-used word classification network by generating a training set through rarely-used words plus culture and noise comprises:
coding each word in the complex word dictionary and the uncommon word dictionary, making a picture corresponding to each corresponding word by using fonts of different styles based on the coding result, and adding noise into the picture to enhance the extended picture so as to obtain a training set;
and training the training set by adopting a classification neural network to obtain the rarely-used word classification network.
5. The method for identifying family tree based on deep learning of claim 1, wherein after obtaining the content of the Chinese characters in the family tree image through the rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image, the method further comprises:
extracting relation words between name information and characters in the family tree image through a regular expression based on the position of each Chinese character and the content of the Chinese character in the family tree image;
and drawing a corresponding family tree diagram based on the relation words between the name information and the characters.
6. A family tree recognition system based on deep learning is characterized by comprising:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring a family tree image;
the second module is used for obtaining the position of each Chinese character in the family tree image through a target position detection network based on deep learning;
and the third module is used for obtaining the content of the Chinese characters in the family tree image through a rarely-used character classification network based on deep learning and the position of each Chinese character in the family tree image.
7. The deep learning based pedigree recognition system of claim 6, wherein the second module is further configured to train the target location detection network through a training set of documents labeled with literal locations;
the third module is also used for generating a training set through the uncommon word plus the culture and the noise to train the uncommon word classification network.
8. The deep learning based pedigree recognition system of claim 7, wherein the second module is configured to train a convolutional neural network as the target location detection network through a set of text data sets with location information.
9. The deep learning-based pedigree recognition system of claim 7, wherein the third module is configured to encode each of the traditional word and the uncommon word dictionary, and based on the encoding result, make a picture corresponding to each of the traditional word and the uncommon word dictionary using different styles of fonts, and add noise to the picture to enhance the extended picture to obtain the training set; and training the training set by adopting a classification neural network to obtain the rarely-used word classification network.
10. The family tree recognition system based on deep learning of claim 6, further comprising a fourth module for extracting, by regular expression, related terms between the name information and the characters in the family tree image based on the position of each character and the content of the characters in the family tree image; and drawing a corresponding family tree diagram based on the relation words between the name information and the characters.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2019106750945 | 2019-07-24 | ||
| CN201910675094 | 2019-07-24 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110852359A true CN110852359A (en) | 2020-02-28 |
| CN110852359B CN110852359B (en) | 2023-05-26 |
Family
ID=69598359
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911035972.3A Active CN110852359B (en) | 2019-07-24 | 2019-10-29 | Family tree identification method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110852359B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111612081A (en) * | 2020-05-25 | 2020-09-01 | 深圳前海微众银行股份有限公司 | Recognition model training method, device, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101025760A (en) * | 2007-01-31 | 2007-08-29 | 王宏源 | Method for digitalizing family tree |
| CN102036018A (en) * | 2009-10-02 | 2011-04-27 | 索尼公司 | Information processing apparatus and method |
| CN107220638A (en) * | 2017-07-03 | 2017-09-29 | 深圳市唯特视科技有限公司 | A kind of car plate detection recognition methods based on deep learning convolutional neural networks |
| CN108664996A (en) * | 2018-04-19 | 2018-10-16 | 厦门大学 | A kind of ancient writing recognition methods and system based on deep learning |
| US20190147304A1 (en) * | 2017-11-14 | 2019-05-16 | Adobe Inc. | Font recognition by dynamically weighting multiple deep learning neural networks |
| CN109800437A (en) * | 2019-01-31 | 2019-05-24 | 北京工业大学 | A kind of name entity recognition method based on Fusion Features |
-
2019
- 2019-10-29 CN CN201911035972.3A patent/CN110852359B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101025760A (en) * | 2007-01-31 | 2007-08-29 | 王宏源 | Method for digitalizing family tree |
| CN102036018A (en) * | 2009-10-02 | 2011-04-27 | 索尼公司 | Information processing apparatus and method |
| CN107220638A (en) * | 2017-07-03 | 2017-09-29 | 深圳市唯特视科技有限公司 | A kind of car plate detection recognition methods based on deep learning convolutional neural networks |
| US20190147304A1 (en) * | 2017-11-14 | 2019-05-16 | Adobe Inc. | Font recognition by dynamically weighting multiple deep learning neural networks |
| CN108664996A (en) * | 2018-04-19 | 2018-10-16 | 厦门大学 | A kind of ancient writing recognition methods and system based on deep learning |
| CN109800437A (en) * | 2019-01-31 | 2019-05-24 | 北京工业大学 | A kind of name entity recognition method based on Fusion Features |
Non-Patent Citations (2)
| Title |
|---|
| MOHAMMED ALMUASHI 等: "Automated kinship verification and identification through human facial images: a survey", MULTIMEDIA TOOLS AND APPLICATIONS * |
| 徐丰;王海鹏;金亚秋: "深度学习在SAR目标识别与地物分类中的应用", 《雷达学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111612081A (en) * | 2020-05-25 | 2020-09-01 | 深圳前海微众银行股份有限公司 | Recognition model training method, device, equipment and storage medium |
| CN111612081B (en) * | 2020-05-25 | 2024-04-02 | 深圳前海微众银行股份有限公司 | Training method, device, equipment and storage medium for recognition model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110852359B (en) | 2023-05-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108664996B (en) | Ancient character recognition method and system based on deep learning | |
| US7310773B2 (en) | Removal of extraneous text from electronic documents | |
| WO2007086059A2 (en) | Determining near duplicate 'noisy' data objects | |
| JP2019502979A (en) | Automatic interpretation of structured multi-field file layouts | |
| CN106372053B (en) | Syntactic analysis method and device | |
| CN106650664A (en) | Collection system and method for college and university enrollment book data | |
| US9558400B2 (en) | Search by stroke | |
| CN113901933A (en) | Electronic invoice information extraction method, device and equipment based on artificial intelligence | |
| CN111104159A (en) | Annotation positioning method based on program analysis and neural network | |
| CN1342942A (en) | Computer recognizing and indexing method of Chinese names | |
| CN109670365B (en) | Handwriting identification system and method | |
| CN105404903B (en) | Information processing method and device and electronic equipment | |
| CN110852359A (en) | Family tree identification method and system based on deep learning | |
| WO2007070010A1 (en) | Improvements in electronic document analysis | |
| KR100655916B1 (en) | Document image processing and verification system for digitalizing a large volume of data and method thereof | |
| CN115830620A (en) | Archive text data processing method and system based on OCR | |
| CN108733733B (en) | Biomedical text classification method, system and storage medium based on machine learning | |
| CN1029534C (en) | Handwriting Chinese character online identifying method and system | |
| CN112560849B (en) | Neural network algorithm-based grammar segmentation method and system | |
| CN109740097A (en) | A kind of Web page text extracting method of logic-based chained block | |
| CN115438645A (en) | Text data enhancement method and system for sequence labeling task | |
| CN1955979A (en) | Automatic extraction device, method and program of essay title and correlation information | |
| CN107368472B (en) | Storage method of document analysis result capable of being iteratively optimized | |
| CN112905733A (en) | Book storage method, system and device based on OCR recognition technology | |
| CN1916888A (en) | Method and system of identifying language of double-byte character set character data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |