CN111401312A - PDF drawing character recognition method, system and equipment - Google Patents
PDF drawing character recognition method, system and equipment Download PDFInfo
- Publication number
- CN111401312A CN111401312A CN202010278085.5A CN202010278085A CN111401312A CN 111401312 A CN111401312 A CN 111401312A CN 202010278085 A CN202010278085 A CN 202010278085A CN 111401312 A CN111401312 A CN 111401312A
- Authority
- CN
- China
- Prior art keywords
- image
- characters
- region
- processing
- 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
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000001514 detection method Methods 0.000 claims abstract description 25
- 238000012015 optical character recognition Methods 0.000 claims abstract description 19
- 238000013135 deep learning Methods 0.000 claims abstract description 10
- 102100032202 Cornulin Human genes 0.000 claims abstract description 9
- 101000920981 Homo sapiens Cornulin Proteins 0.000 claims abstract description 9
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 9
- 238000013528 artificial neural network Methods 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims description 56
- 238000000605 extraction Methods 0.000 claims description 40
- 238000004590 computer program Methods 0.000 claims description 15
- 238000004458 analytical method Methods 0.000 claims description 13
- 239000000284 extract Substances 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 8
- 238000003709 image segmentation Methods 0.000 claims description 6
- 230000000877 morphologic effect Effects 0.000 claims description 6
- 238000003706 image smoothing Methods 0.000 claims description 3
- 230000010339 dilation Effects 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 5
- 235000019587 texture Nutrition 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 235000019580 granularity Nutrition 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000009432 framing Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000011282 treatment Methods 0.000 description 1
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
- G06V30/40—Document-oriented image-based pattern recognition
- G06V30/41—Analysis of document content
- G06V30/414—Extracting the geometrical structure, e.g. layout tree; Block segmentation, e.g. bounding boxes for graphics or text
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
-
- 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
- G06V30/40—Document-oriented image-based pattern recognition
- G06V30/42—Document-oriented image-based pattern recognition based on the type of document
- G06V30/422—Technical drawings; Geographical maps
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Health & Medical Sciences (AREA)
- Character Discrimination (AREA)
- Character Input (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a PDF drawing character recognition method, a system and equipment, wherein the PDF drawing character recognition method comprises the following steps: performing an optical character recognition step based on the deep learning; customized identification and universal identification; and a mobile device low quality image recognition step; wherein the step of performing optical character recognition based on deep learning includes the steps of: detecting a region with characters in a scene and identifying the characters in the region, wherein text detection is performed based on CTPN, Seglink, TextBox, FTSN, Pixellink and CRAFT algorithms; the characters are identified based on CNN and CRNN algorithms; wherein the step of customizing the identification comprises the steps of: identifying the type of a PDF drawing according to the table characters in the PDF or the frame content in the PDF; extracting content in the region according to the structural features; and extracting a key area, and identifying characters in the area or extracting key characters through a deep neural network.
Description
Technical Field
The invention relates to the field of image processing, in particular to a method, a system and equipment for identifying characters of a PDF drawing.
Background
The artificial intelligence has made rapid development in the aspects of data, algorithm and computing power, and a new round of development wave is met under the large background of global economy digital transformation. The influence of the artificial intelligence wave is far beyond the prior art, and the most remarkable characteristic is that the influence is diffused from the professional field to the popular field.
PDF high-precision identification is a rather mature technology in the current market, and a method based on traditional OCR and deep learning is also applied to various industries. Identification of bank notes, PDF forms and industrial drawings are all widely and mature technologies that have been used. Identification of formatted and templated PDFs has dramatic results in both accuracy and speed, thereby improving the work efficiency and capacity of practitioners in various industries.
The conventional PDF identifies PDF in a basic fixed form, and has certain requirements on PDF quality. With the popularization of smart phones, the traditional method does not have a good solution for low-quality PDF images shot by personal mobile phones.
Most of the current PDF identification is whole identification, and extraction and identification aiming at special areas are not provided. For some structured PDF drawings, extracting a region (POI) which is interested by a user and analyzing the content in the POI is also a characteristic of the scheme.
Disclosure of Invention
The invention aims to provide a method, a system and equipment for recognizing characters on a PDF drawing, which provide various recognition schemes such as customized and general scenes, and can provide a solution for recognizing low-quality images by a user based on a deep learning nearest OCR algorithm and a corresponding image processing technology.
Another object of the present invention is to provide a method, a system and a device for identifying characters on a PDF drawing, which can identify various scenes such as industrial drawings, bills, images shot by personal devices, and solve different requirements of different users.
In order to achieve at least one of the above objects, the present invention provides a method for identifying characters on a PDF drawing, wherein the method for identifying characters on a PDF drawing comprises the following steps:
performing an optical character recognition step based on the deep learning; customized identification and universal identification; and a mobile device low quality image recognition step;
wherein the performing optical character recognition based on deep learning includes: detecting a region with characters in a scene and identifying the characters in the region, wherein text detection is performed based on CTPN, Seglink, TextBox, FTSN, Pixellink and CRAFT algorithms; the characters are identified based on CNN and CRNN algorithms;
wherein the step of customizing identification comprises the steps of: identifying the type of a PDF drawing according to the table characters in the PDF or the frame content in the PDF; extracting content in the region according to the structural features; and extracting a key area, and identifying characters in the area or extracting key characters through a deep neural network.
In some embodiments, the step of extracting key regions of the step of customizing identification further comprises the steps of:
extracting a key area according to the proportion of the POI;
extracting all frames by using Hough transform and corner detection;
extracting a key area according to fuzzy matching and accurate matching of characters;
extracting a key area according to the edge characteristics of the area; and
and extracting a key area according to the character characteristics in the area.
In some embodiments, in the step of extracting key regions in the step of customizing identification, key regions are extracted according to edge characteristics of shapes or symmetries or angles or edge granularities of the regions, wherein the key regions are extracted according to characteristics of fonts or sizes of characters in the regions or character types.
In some embodiments, wherein the mobile device low quality image recognition step further comprises the steps of:
performing filtering processing on the image;
performing image enhancement processing on the image;
performing image marginalization sharpening processing on the image;
performing image texture analysis processing on the image;
performing image segmentation processing on the image;
performing geometric analysis processing on the image;
performing image matching processing on the image; and
morphological processing is performed on the image.
In some embodiments, the filtering the image comprises performing image smoothing and image denoising on the image; wherein the performing image texture analysis processing on the image is performing skeleton removal and connectivity processing on the image; wherein the step of performing image matching processing on the image is to perform template matching and search matching processing on the image; wherein the step of performing morphological processing on the image is performing dilation, apparel, and opening and closing operations on the image.
According to another aspect of the present invention, a computer-readable storage medium is further provided, on which a computer program is stored, which, when being executed by a processor, performs the steps of the PDF drawing text recognition method.
According to another aspect of the present invention, there is also provided a PDF drawing text recognition apparatus, including: a software application, a memory for storing the software application, and a processor for executing the software application; and correspondingly executing the steps in the PDF drawing character recognition method by each program of the software application program.
According to another aspect of the present invention, there is also provided a PDF drawing text recognition system, comprising an optical character recognition unit, a customized recognition and general purpose recognition unit, and a mobile device low-quality image recognition unit, wherein the optical character recognition unit comprises a text detection module and a text recognition module, wherein the text detection module is configured to: detecting a region with characters in a scene, and executing text detection based on CTPN, Seglink, TextBox, FTSN, Pixellink and CRAFT algorithms; wherein the word recognition module is configured to: recognizing characters in the detected region, and recognizing the characters based on CRNN and CNN algorithms; the customized identification and general identification unit is provided with a customized identification module, the customized identification module is configured to structurally extract the content in the region, and identify characters in the region or extract key characters through a deep neural network; wherein the mobile device low quality image recognition unit is configured to: performing filtering processing, image enhancement processing, image marginalization sharpening processing, image texture analysis processing, image segmentation processing, geometric shape analysis processing, image matching processing, and morphological processing on the identified image
In some embodiments, the key region extraction module further comprises a POI scale extraction module, a hough transform corner detection extraction module, a text blurring and exact matching extraction module, a region edge characteristic extraction module, and a region text characteristic extraction module; wherein the POI scale extraction module is configured to extract key regions according to POI scale size; the Hough transform corner detection extraction module is configured to extract all frames by utilizing Hough transform and corner detection; the character fuzzy and precise matching extraction module is configured to extract a key area according to fuzzy matching and precise matching of characters; the region edge characteristic extraction module is configured to extract a key region according to edge characteristics of the region; the regional text characteristic extraction module is configured to extract key regions according to text characteristics in the regions.
In some embodiments, the edge characteristics of the region edge characteristic extraction module are shape, symmetry, angle and edge granularity, and the text characteristics in the region of the region text characteristic extraction module are font, size and text type.
Drawings
Fig. 1 is a flowchart illustrating steps of a method for recognizing characters on a PDF drawing according to an embodiment of the present invention.
Fig. 2 is a schematic structural diagram of a PDF drawing text recognition system according to an embodiment of the present invention.
Detailed Description
The following description is presented to disclose the invention so as to enable any person skilled in the art to practice the invention. The preferred embodiments in the following description are given by way of example only, and other obvious variations will occur to those skilled in the art. The basic principles of the invention, as defined in the following description, may be applied to other embodiments, variations, modifications, equivalents, and other technical solutions without departing from the spirit and scope of the invention.
It is understood that the terms "a" and "an" should be interpreted as meaning that a number of one element or element is one in one embodiment, while a number of other elements is one in another embodiment, and the terms "a" and "an" should not be interpreted as limiting the number.
The present invention relates to a computer program. Fig. 1 is a flow chart of a method for recognizing characters on a PDF drawing according to the present invention, which illustrates a solution for controlling or processing an external object or an internal object of a computer by executing a computer program compiled according to the above flow on the basis of a computer program processing flow to solve the problems of the present invention. By the PDF drawing character recognition method, a computer system can be utilized, artificial experience and machine learning results can be integrated, recognition of various scenes such as industrial drawings, bills and images shot by personal equipment can be achieved, and different requirements of different users can be met.
As shown in fig. 1, the method for recognizing characters on a PDF drawing includes the following steps:
s10: an optical character recognition step is performed based on the deep learning.
The optical character recognition step includes the steps of:
detecting a region with characters in a scene, wherein text detection is executed based on CTPN, Seglink, TextBoxes, FTSN, Pixellink and CRAFT algorithms; and
and recognizing characters in the detected region, wherein the characters are recognized based on the CRNN algorithm and the CNN algorithm.
Further, the method for recognizing characters on a PDF drawing further includes step S20: customized identification and universal identification.
The customized identification and universal identification steps comprise the following steps:
customizing and identifying; and
and (5) universal identification.
Wherein the step of customizing the identification further comprises the steps of:
identifying the type of a PDF drawing according to the table characters in the PDF or the frame content in the PDF;
extracting content in the region according to the structural features; and
and extracting a key area, and identifying characters in the area or extracting key characters through a deep neural network.
Wherein the step of extracting the key area further comprises the following steps:
extracting according to the proportion of POI;
extracting all frames by using Hough transform and corner detection;
fuzzy matching and accurate matching of characters;
according to the edge characteristics of the region, such as shape, symmetry, angle, edge granularity, and the like; and
depending on the character properties in the region, such as font, size, character type, etc.
Further, the method for recognizing characters on a PDF drawing further includes step S30: a mobile device low quality image recognition step.
The mobile device low quality image recognition step further comprises the steps of:
filtering, e.g., image smoothing, image denoising;
enhancing the image;
sharpening the edge of the image;
image texture analysis, e.g., de-framing, connectivity;
image segmentation;
analyzing the geometrical morphology;
image matching, e.g., template matching, search matching; and
morphological treatments such as inflation, apparel, opening and closing operations, and the like.
The enterprise provides an identification entrance, and the enterprise is provided with the capability of shooting PDF drawings and identifying individual users. However, the image taken by the user is often poor in quality due to the lighting condition, the shooting angle, and the like. The PDF drawing character recognition method can improve the quality of images, thereby improving the recognition precision. Aiming at the low-quality image provided by the personal equipment provided by the user, after the low-quality image recognition step of the mobile equipment is carried out, the image can have better expressive force, and the image quality approaches to the high-precision PDF, so that the recognition algorithm can be better carried out. But also can expand the applicable scenes and generalization of the whole algorithm.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects.
It can be understood by those skilled in the art that the method for identifying the text on the PDF drawing according to the present invention can be implemented by hardware, software, or a combination of hardware and software. The present invention can be realized in a centralized fashion in at least one computer system, or in a distributed fashion where different elements are spread across several interconnected computer systems. Any kind of computer system or other apparatus adapted for carrying out the methods described herein is suited. A typical combination of hardware and software could be a general purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein.
The present invention can be embedded in a computer program product, which comprises all the features enabling the implementation of the methods described herein. The computer program product is embodied in one or more computer-readable storage media having computer-readable program code embodied therein. According to another aspect of the invention, there is also provided a computer-readable storage medium having stored thereon a computer program capable, when executed by a processor, of performing the steps of the method of the invention. Computer storage media is media in computer memory for storage of some discrete physical quantity. Computer storage media includes, but is not limited to, semiconductors, magnetic disk storage, magnetic cores, magnetic drums, magnetic tape, laser disks, and the like. It will be appreciated by persons skilled in the art that computer storage media are not limited by the foregoing examples, which are intended to be illustrative only and not limiting of the invention.
A typical combination of hardware and software could be a general purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein. According to another aspect of the present invention, there is also provided a PDF drawing text recognition apparatus, including: a software application, a memory for storing the software application, and a processor for executing the software application. And each program of the software application program can correspondingly execute the steps in the PDF drawing character recognition method.
Corresponding to the embodiment of the method, according to another aspect of the invention, a PDF drawing character recognition system is also provided, and the PDF drawing character recognition system is an application of the PDF drawing character recognition method in the improvement of computer programs.
As shown in fig. 2, in this embodiment of the present invention, the PDF drawing text recognition system includes an optical character recognition unit 100, a customized recognition and general purpose recognition unit 200, and a mobile device low-quality image recognition unit 300.
Specifically, the optical character recognition unit 100 includes a text detection module 110 and a recognition module 120. Wherein the text detection module 110 detects a region with words in the scene, preferably, text detection is performed based on CTPN, Seglink, TextBoxes, FTSN, Pixellink, and CRAFT algorithms. The character recognition module 120 recognizes characters in the detected region, preferably, recognizes characters based on CRNN and CNN algorithms. It will be appreciated by those skilled in the art that in other embodiments of the invention, algorithms other than CRNN, CNN, etc. may be used, and the invention is not limited in this respect.
With the rapid development of neural networks in computer vision, the precision of Optical Character Recognition (OCR) is also greatly improved compared with the conventional technology. Under the large background of deep learning, character recognition is also expanded from the recognition of traditional scenes to the character recognition of general scenes, namely, the character recognition of natural scenes. The algorithm model of the optical character recognition unit ensures that the recognition accuracy of the user is ensured to the maximum extent. In addition, the speed and the precision may be different for different users, so according to different scenes and users, the PDF drawing text recognition system adapts to multiple algorithms, and the text recognition module 120 replaces some complex text recognition algorithms (such as CRNN) with CNN, which can provide a recognition scheme with a faster rate as possible for simple scenes.
The customized identification and general identification unit 200 is provided with a customized identification module 210, and the customized identification module 210 is configured to structure the content in the feature extraction area and identify the words in the area or extract key words through a deep neural network. Preferably, the customized identification module 210 is provided with a key region extraction module 220, and the key region extraction module 220 further includes a POI proportion extraction module 221, a hough transformation corner detection extraction module 222, a text blurring and exact matching extraction module 223, a region edge characteristic extraction module 224, and a region text characteristic extraction module 225. The POI proportion extraction module 221 extracts according to the POI proportion; the hough transform corner detection and extraction module 222 extracts all frames by using hough transform and corner detection; the text fuzzy and precise matching extraction module 223 extracts the content in the key area according to the fuzzy matching and precise matching of the text; the region edge characteristic extraction module 224 extracts contents in the key region, such as shape, symmetry, angle, edge granularity, and the like, according to the edge characteristics of the region); the regional text property extraction module 225 extracts content, such as font, size, text type, etc., in the key region according to the text properties in the region.
The mobile device low-quality image recognition unit 300 is configured to perform filtering processing, image enhancement processing, image marginalization sharpening processing, image texture analysis processing, image segmentation processing, geometric shape analysis processing, image matching processing, and morphology processing on the recognized image.
The enterprise provides an identification entrance, and the enterprise is provided with the capability of shooting PDF drawings and identifying individual users. However, the image taken by the user is often poor in quality due to the lighting condition, the shooting angle, and the like. Aiming at the problems, the low-quality image recognition unit 300 of the mobile equipment adopts a plurality of image processing schemes, so that the quality of the image is improved, and the recognition accuracy is improved.
Aiming at low-quality images provided by personal equipment provided by a user, after the low-quality images are processed by the PDF drawing character recognition system, the images can have better expressive force, and the image quality approaches to high-precision PDF, so that the recognition algorithm can be better carried out, and the application scene and the generalization of the whole algorithm can be expanded.
It will be appreciated by those skilled in the art that the present invention has been described with reference to flowchart illustrations and/or block diagrams of methods, systems and computer program products according to the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and/or block diagram block or blocks.
It will be appreciated by persons skilled in the art that the embodiments of the invention described above and shown in the drawings are given by way of example only and are not limiting of the invention. The objects of the invention have been fully and effectively accomplished. The functional and structural principles of the present invention have been shown and described in the examples, and any variations or modifications of the embodiments of the present invention may be made without departing from the principles.
Claims (10)
1. A PDF drawing character recognition method is characterized by comprising the following steps:
performing an optical character recognition step based on the deep learning; customized identification and universal identification; and a mobile device low quality image recognition step;
wherein the performing optical character recognition based on deep learning includes: detecting a region with characters in a scene and identifying the characters in the region, wherein text detection is performed based on CTPN, Seglink, TextBox, FTSN, Pixellink and CRAFT algorithms; the characters are identified based on CNN and CRNN algorithms;
wherein the step of customizing identification comprises the steps of: identifying the type of a PDF drawing according to the table characters in the PDF or the frame content in the PDF; extracting content in the region according to the structural features; and extracting a key area, and identifying characters in the area or extracting key characters through a deep neural network.
2. The method for recognizing characters on PDF drawing as recited in claim 1, wherein said step of extracting key regions in said step of customizing recognition further comprises the steps of:
extracting a key area according to the proportion of the POI;
extracting all frames by using Hough transform and corner detection;
extracting a key area according to fuzzy matching and accurate matching of characters;
extracting a key area according to the edge characteristics of the area; and
and extracting a key area according to the character characteristics in the area.
3. The method for recognizing characters on PDF drawing as claimed in claim 2, wherein in the step of extracting key regions in the customized recognition step, key regions are extracted according to the edge characteristics of the shapes or symmetries or angles or edge granularity of the regions, wherein the key regions are extracted according to the characters fonts or sizes or characters types in the regions.
4. The method of claim 1, wherein the mobile device low quality image recognition step further comprises the steps of:
performing filtering processing on the image;
performing image enhancement processing on the image;
performing image marginalization sharpening processing on the image;
performing image texture analysis processing on the image;
performing image segmentation processing on the image;
performing geometric analysis processing on the image;
performing image matching processing on the image; and
morphological processing is performed on the image.
5. The method for recognizing characters on PDF drawing as recited in claim 4, wherein said step of performing filtering process on image includes performing image smoothing and image denoising process on image; wherein the performing image texture analysis processing on the image is performing skeleton removal and connectivity processing on the image; wherein the step of performing image matching processing on the image is to perform template matching and search matching processing on the image; wherein the step of performing morphological processing on the image is performing dilation, apparel, and opening and closing operations on the image.
6. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the PDF drawing text recognition method according to any of claims 1 to 5.
7. The utility model provides a PDF drawing characters recognition equipment which characterized in that, PDF drawing characters recognition equipment includes: a software application, a memory for storing the software application, and a processor for executing the software application; the method for identifying characters on PDF drawings is characterized in that each program of the software application program correspondingly executes the steps in the method for identifying characters on PDF drawings as claimed in any one of claims 1 to 5.
8. A PDF drawing text recognition system, comprising an optical character recognition unit, a customized recognition and generic recognition unit, and a mobile device low quality image recognition unit, wherein the optical character recognition unit comprises a text detection module and a text recognition module, wherein the text detection module is configured to: detecting a region with characters in a scene, and executing text detection based on CTPN, Seglink, TextBox, FTSN, Pixellink and CRAFT algorithms; wherein the word recognition module is configured to: recognizing characters in the detected region, and recognizing the characters based on CRNN and CNN algorithms; the customized identification and general identification unit is provided with a customized identification module, the customized identification module is configured to structurally extract the content in the region, and identify characters in the region or extract key characters through a deep neural network; wherein the mobile device low quality image recognition unit is configured to: and carrying out filtering processing, image enhancement processing, image marginalization sharpening processing, image texture analysis processing, image segmentation processing, geometric morphology analysis processing, image matching processing and morphology processing on the identified image.
9. The PDF drawing text recognition system of claim 8, wherein the key region extraction module further comprises a POI scale extraction module, a hough transform corner detection extraction module, a text blur and exact match extraction module, a region edge characteristic extraction module, and a region text characteristic extraction module; wherein the POI scale extraction module is configured to extract key regions according to POI scale size; the Hough transform corner detection extraction module is configured to extract all frames by utilizing Hough transform and corner detection; the character fuzzy and precise matching extraction module is configured to extract a key area according to fuzzy matching and precise matching of characters; the region edge characteristic extraction module is configured to extract a key region according to edge characteristics of the region; the regional text characteristic extraction module is configured to extract key regions according to text characteristics in the regions.
10. The PDF drawing text recognition system of claim 9, wherein the edge characteristics of the region edge characteristic extraction module are shape, symmetry, angle and edge granularity, and wherein the text characteristics in the region of the region text characteristic extraction module are font, size and text type.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010278085.5A CN111401312B (en) | 2020-04-10 | 2020-04-10 | PDF drawing text recognition method, system and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010278085.5A CN111401312B (en) | 2020-04-10 | 2020-04-10 | PDF drawing text recognition method, system and equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111401312A true CN111401312A (en) | 2020-07-10 |
| CN111401312B CN111401312B (en) | 2024-04-26 |
Family
ID=71435039
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010278085.5A Active CN111401312B (en) | 2020-04-10 | 2020-04-10 | PDF drawing text recognition method, system and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111401312B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814791A (en) * | 2020-07-24 | 2020-10-23 | 西门子（中国）有限公司 | Method and device for identifying components in system diagram |
| CN112183029A (en) * | 2020-09-25 | 2021-01-05 | 四川巧夺天工信息安全智能设备有限公司 | Digital conversion method for PDF drawing in sheet metal industry |
| CN112733735A (en) * | 2021-01-13 | 2021-04-30 | 国网上海市电力公司 | Method for classifying and identifying drawing layout by machine learning |
| CN113094786A (en) * | 2021-04-06 | 2021-07-09 | 万翼科技有限公司 | Construction drawing structured organization method and device based on drawing POI |
| CN113743052A (en) * | 2021-08-17 | 2021-12-03 | 的卢技术有限公司 | Multi-mode-fused resume layout analysis method and device |
| CN114170616A (en) * | 2021-11-15 | 2022-03-11 | 嵊州市光宇实业有限公司 | Electric power engineering material information acquisition and analysis system and method based on graph paper set |
| CN117558019A (en) * | 2024-01-11 | 2024-02-13 | 武汉理工大学 | Method for automatically extracting symbol map parameters from PDF format component manual |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050289182A1 (en) * | 2004-06-15 | 2005-12-29 | Sand Hill Systems Inc. | Document management system with enhanced intelligent document recognition capabilities |
| US20190050639A1 (en) * | 2017-08-09 | 2019-02-14 | Open Text Sa Ulc | Systems and methods for generating and using semantic images in deep learning for classification and data extraction |
| CN110291536A (en) * | 2017-01-30 | 2019-09-27 | 赛门铁克公司 | For preventing the structured text and pattern match of the loss of data in object specific pattern image field |
| CN110334346A (en) * | 2019-06-26 | 2019-10-15 | 京东数字科技控股有限公司 | A kind of information extraction method and device of pdf document |
| CN110363102A (en) * | 2019-06-24 | 2019-10-22 | 北京融汇金信信息技术有限公司 | A kind of identification of objects process method and device of pdf document |
| CN110543844A (en) * | 2019-08-26 | 2019-12-06 | 中电科大数据研究院有限公司 | metadata extraction method for government affair metadata PDF file |
| CN110569769A (en) * | 2019-08-29 | 2019-12-13 | 浙江大搜车软件技术有限公司 | image recognition method and device, computer equipment and storage medium |
| CN110751143A (en) * | 2019-09-26 | 2020-02-04 | 中电万维信息技术有限责任公司 | Electronic invoice information extraction method and electronic equipment |
-
2020
- 2020-04-10 CN CN202010278085.5A patent/CN111401312B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050289182A1 (en) * | 2004-06-15 | 2005-12-29 | Sand Hill Systems Inc. | Document management system with enhanced intelligent document recognition capabilities |
| CN110291536A (en) * | 2017-01-30 | 2019-09-27 | 赛门铁克公司 | For preventing the structured text and pattern match of the loss of data in object specific pattern image field |
| US20190050639A1 (en) * | 2017-08-09 | 2019-02-14 | Open Text Sa Ulc | Systems and methods for generating and using semantic images in deep learning for classification and data extraction |
| CN110363102A (en) * | 2019-06-24 | 2019-10-22 | 北京融汇金信信息技术有限公司 | A kind of identification of objects process method and device of pdf document |
| CN110334346A (en) * | 2019-06-26 | 2019-10-15 | 京东数字科技控股有限公司 | A kind of information extraction method and device of pdf document |
| CN110543844A (en) * | 2019-08-26 | 2019-12-06 | 中电科大数据研究院有限公司 | metadata extraction method for government affair metadata PDF file |
| CN110569769A (en) * | 2019-08-29 | 2019-12-13 | 浙江大搜车软件技术有限公司 | image recognition method and device, computer equipment and storage medium |
| CN110751143A (en) * | 2019-09-26 | 2020-02-04 | 中电万维信息技术有限责任公司 | Electronic invoice information extraction method and electronic equipment |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814791A (en) * | 2020-07-24 | 2020-10-23 | 西门子（中国）有限公司 | Method and device for identifying components in system diagram |
| CN111814791B (en) * | 2020-07-24 | 2024-03-19 | 西门子（中国）有限公司 | Method and device for identifying components in system graph |
| CN112183029A (en) * | 2020-09-25 | 2021-01-05 | 四川巧夺天工信息安全智能设备有限公司 | Digital conversion method for PDF drawing in sheet metal industry |
| CN112183029B (en) * | 2020-09-25 | 2023-10-31 | 四川巧夺天工信息安全智能设备有限公司 | Digital conversion method for PDF drawing in sheet metal industry |
| CN112733735A (en) * | 2021-01-13 | 2021-04-30 | 国网上海市电力公司 | Method for classifying and identifying drawing layout by machine learning |
| CN112733735B (en) * | 2021-01-13 | 2024-04-09 | 国网上海市电力公司 | Method for classifying and identifying drawing layout by adopting machine learning |
| CN113094786A (en) * | 2021-04-06 | 2021-07-09 | 万翼科技有限公司 | Construction drawing structured organization method and device based on drawing POI |
| CN113743052A (en) * | 2021-08-17 | 2021-12-03 | 的卢技术有限公司 | Multi-mode-fused resume layout analysis method and device |
| CN113743052B (en) * | 2021-08-17 | 2024-06-14 | 的卢技术有限公司 | Method and device for analyzing layout of resume fused with multiple modes |
| CN114170616A (en) * | 2021-11-15 | 2022-03-11 | 嵊州市光宇实业有限公司 | Electric power engineering material information acquisition and analysis system and method based on graph paper set |
| CN117558019A (en) * | 2024-01-11 | 2024-02-13 | 武汉理工大学 | Method for automatically extracting symbol map parameters from PDF format component manual |
| CN117558019B (en) * | 2024-01-11 | 2024-05-28 | 武汉理工大学 | Method for automatically extracting symbol map parameters from PDF format component manual |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111401312B (en) | 2024-04-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111401312B (en) | PDF drawing text recognition method, system and equipment | |
| Saxena | Niblack’s binarization method and its modifications to real-time applications: a review | |
| CN114359998B (en) | Identification method of face mask in wearing state | |
| CN103208004A (en) | Automatic recognition and extraction method and device for bill information area | |
| Ardizzone et al. | Saliency based image cropping | |
| CN103902977A (en) | Face identification method and device based on Gabor binary mode | |
| Zheng et al. | A cascaded method for text detection in natural scene images | |
| Bai et al. | Scene text localization using gradient local correlation | |
| CN106127193B (en) | A kind of facial image recognition method | |
| CN117197904B (en) | Training method of human face living body detection model, human face living body detection method and human face living body detection device | |
| Huang et al. | Text detection and recognition in natural scene images | |
| CN109086772A (en) | A kind of recognition methods and system distorting adhesion character picture validation code | |
| Meng et al. | Ancient Asian character recognition for literature preservation and understanding | |
| Lahiani et al. | Hand pose estimation system based on Viola-Jones algorithm for android devices | |
| CN112686265A (en) | Hierarchic contour extraction-based pictograph segmentation method | |
| Li et al. | Leveraging surrounding context for scene text detection | |
| Ramzi et al. | Online Arabic handwritten character recognition using online-offline feature extraction and back-propagation neural network | |
| Wicht et al. | Camera-based sudoku recognition with deep belief network | |
| Lin et al. | A traffic sign recognition method based on deep visual feature | |
| Azad et al. | Real-time human face detection in noisy images based on skin color fusion model and eye detection | |
| Wenhong et al. | Application of support vector machine (SVM) on serial number identification of RMB | |
| CN112101293A (en) | Facial expression recognition method, device, equipment and storage medium | |
| Ali et al. | A new design based-fusion of features to recognize Arabic handwritten characters | |
| Thilagavathy et al. | Fuzzy based edge enhanced text detection algorithm using MSER | |
| Chatbri et al. | Shape matching using keypoints extracted from both the foreground and the background of binary images |
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 |