CN110298354A - A kind of facility information identifying system and its recognition methods - Google Patents
A kind of facility information identifying system and its recognition methods Download PDFInfo
- Publication number
- CN110298354A CN110298354A CN201910611848.0A CN201910611848A CN110298354A CN 110298354 A CN110298354 A CN 110298354A CN 201910611848 A CN201910611848 A CN 201910611848A CN 110298354 A CN110298354 A CN 110298354A
- Authority
- CN
- China
- Prior art keywords
- information
- character
- module
- character recognition
- processing module
- 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
Classifications
-
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/146—Aligning or centring of the image pick-up or image-field
- G06V30/1475—Inclination or skew detection or correction of characters or of image to be recognised
- G06V30/1478—Inclination or skew detection or correction of characters or of image to be recognised of characters or characters lines
-
- 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
-
- 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)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Character Discrimination (AREA)
Abstract
The present invention proposes a kind of facility information identifying system and its recognition methods.The system includes, image pre-processing module, character recognition and correction module, data processing module, data memory module, described image preprocessing module receives the information of image collection module transmission and is transmitted to character recognition and correction module by its binary conversion treatment and by the information after binary conversion treatment, the character recognition and correction module are to identify the information and the corrected information of character recognition is transmitted to data processing module, information and pre-stored information of the data processing module based on the character recognition and correction module transmission are matched, and matched information is transmitted to data memory module.The name plate information for being carried out intelligent recognition Medical Devices using the system is completed the extraction and typing work of data, greatly improves equipment efficiency of inputting, shorten the investigation period of Medical Devices, and improve data accuracy.
Description
Technical field
The present invention relates to a kind of information identification system, it is specifically related to a kind of medical equipment information identifying system and its identification
Method.
Background technique
In to Medical Devices troubleshooting procedure, need device name in the nameplate every equipment, model, sequence number,
Producer, the date of production and registration certificate number is full and accurate is recorded in the system of Medical Devices, so as to by software to medical treatment
Equipment carries out file administration, maintenance, maintenance etc..Current facility information typing, be by manually taking pictures to equipment nameplate after,
The information in photo is transcribed into system again.The embodiment has following deficiency,
(1) efficiency of inputting is low;Artificial control picture logging data, relatively time-consuming, especially sequence number, registration certificate number etc. compare
Longer information, the data inputting of average each nameplate need 1 minute or so time.
(2) the investigation period is long;Since manual entry holding time is long, so that the investigation of all devices of hospital total period becomes
Long, the use of Managerial System of Medical Equipment must be based on most basic device data again, so will affect system use.
(3) it is easy error;For a long time frequently using naked eyes identification nameplate picture, eyes are easily tired, are easy to appear mistake
Accidentally.
Summary of the invention
To solve the problems, such as above-mentioned at least one, the present invention propose a kind of Medical Devices identification information identifying system and its
Recognition methods.It reduces workload artificial when Medical Devices investigation, improves investigation precision.
In order to achieve the above purpose, the application adopts the following technical scheme that
A kind of facility information identifying system, which is characterized in that include image pre-processing module, character recognition and straightening die
Block, data processing module, data memory module,
Described image preprocessing module receives the information of image collection module transmission and by its binary conversion treatment and by two-value
Change that treated that information is transmitted to character recognition and correction module,
The character recognition and correction module are to identify the information and be transmitted to the corrected information of character recognition
Data processing module,
The information and pre-stored information that the data processing module is transmitted based on the character recognition and correction module into
Row matching, and matched information is transmitted to data memory module.
Preferably, which connects database to obtain its pre-stored facility information.
Preferably, the character recognition and correction module arrange the character of identification according to preset rules.
Preferably, the data memory module is certain regularly arranged to the character foundation that will be identified.
Preferably, which also includes the segmentation for going inclination, character picture to the character picture of acquisition
Processing.
Preferably, which is included and is determined to carry out binaryzation to character picture based on global the threshold method of the maximum equation difference
Gray threshold K.
Preferably, which pre-processes to the information to equipment, it includes to character picture into
Row goes inclination to handle.
Preferably, the character recognition and correction module are to the identification and arrangement to candidate characters, with the small echo of character
Energy and edge orientation histogram construct candidate characters decision model with support vector machine classifier as its feature vector,
Optimal alignment is carried out to candidate characters by character arrangements model, obtains final identification information.
The embodiment of the present application also provides a kind of facility information recognition methods, and it includes above-mentioned systems, which is characterized in that institute
The method of stating comprises the following steps:
S1 obtains name plate information based on image processing module,
S2, the name plate information that will acquire are pre-processed,
S3, pretreated information carry out character recognition and arrangement and match it with pre-stored information,
Matched information is carried out regulation storage by S4.
Preferably, in the S2, pretreatment removes inclination and character figure comprising the binaryzation of character picture, character picture
The segmentation of picture.
Preferably, in the S2, also comprising tilt to character picture, the inclination comprising to character picture respectively into
Row x-axis, the projection on y-axis direction, projected length are respectively L1 and L2, then detect the starting point that gray value is 1 in y-axis
Coordinate, the distance to x-axis are denoted as L3, if the tilt angle of going of character picture is θ, then:
Scheme in compared with the existing technology, advantages of the present invention:
Embodiment proposed by the present invention, character recognition and timing on name plate information, that is, nameplate of Medical Devices
Using the wavelet energy of character and edge orientation histogram as its feature vector, candidate word is constructed with support vector machine classifier
Decision model is accorded with, optimal alignment is carried out to candidate characters by character arrangements model, obtains final nameplate recognition result.It makes
With smart machine come intelligent recognition nameplate, the extraction and typing work of data are completed, equipment efficiency of inputting is greatly improved, contracts
The short investigation period of Medical Devices, and improve data accuracy.
Detailed description of the invention
The invention will be further described with reference to the accompanying drawings and embodiments:
Fig. 1 show the functional schematic of the facility information identifying system of the embodiment of the present application.
Fig. 2 show the flow diagram of the facility information recognition methods of the embodiment of the present application.
Specific embodiment
Above scheme is described further below in conjunction with specific embodiment.It should be understood that these embodiments are for illustrating
The present invention and be not limited to limit the scope of the invention.Implementation condition used in the examples can be done such as the condition of specific producer into
One successive step, the implementation condition being not specified are usually the condition in routine experiment.
The embodiment that the application proposes, carrys out intelligent recognition nameplate using smart machine, completes the extraction and record of data
Enter work, greatly improve equipment efficiency of inputting, shortens the investigation period of Medical Devices, and it is accurate to improve data
Rate.
It is as shown in Figure 1 the functional schematic of the facility information identifying system of the embodiment of the present application,
The system includes image pre-processing module, character recognition and correction module, data processing module, and data store mould
Block, the information which receives image collection module transmission are pre-processed, are passed by pretreated information
It transports to character recognition and correction module carries out character recognition correction, and the corrected information of character recognition is transmitted to data processing
Module, the data processing module based on after correction information and pre-stored information matched, and matched information is transmitted
To data memory module.Information of the data memory module to store matched equipment.
In present embodiment, image pre-processing module is to information (e.g., the word on the nameplate of Medical Devices to equipment
Symbol) pre-processed, it includes the binary conversion treatment of character picture, character picture go inclination and character picture segmentation.
The image of equipment nameplate usually has color, for convenience the segmentation and identification to character, needs to carry out two to character picture
Value processing.In one embodiment, the ash that binaryzation is carried out to character picture is determined using global the threshold method of the maximum equation difference
Spend threshold k.Next its treatment process is described, if the gray level range of gray level image is [0, M], a certain gray level K is by the area
Between be divided into two groups, respectively [0, K] and [K+1, M] are denoted as C0 and C1, then between the two gray scale intervals average gray variance
Are as follows:
In formula: μ 0 and μ 1 is respectively the average gray of C0 and C1;ω 0 and ω 1 is respectively pixel quantity shared by C0 and C1
With the ratio of total pixel quantity.The value that K is constantly converted between [0, M], when finding out δ 2 (K) in above formula and being maximum value
K value, this value are exactly the gray threshold K that binaryzation is carried out to character picture.This is arrived, the gray scale of character on equipment nameplate is completed
Binaryzation.
Character when in one embodiment, due to shooting name plate information, due to shooting angle, on equipment nameplate
There are a small angle thetas between meeting and horizontal line.In order to facilitate the extraction of feature, need that character picture incline
Tiltedly.Since the character on nameplate arranges (such as horizontally-arranged writing, i.e., writing from left to right) to character picture according to certain rules
The projection on x-axis and y-axis direction is carried out respectively, and projected length is respectively L1 and L2, and then detecting gray value in y-axis is 1
The coordinate of starting point, the distance to x-axis are denoted as L3, if the tilt angle of going of character picture is θ, then:
For the character on equipment nameplate, if the y-axis that the y-axis projection coordinate of leftmost side character is greater than rightmost side character is thrown
The view field of entire character is then rotated counterclockwise and removes tiltangleθ, tilted conversely, then rotating clockwise by shadow coordinate
Angle θ.After carrying out the binaryzation of character picture and going inclination, followed by the segmentation and normalization of character picture.
The segmentation of equipment nameplate character can be divided between row segmentation and word and divide, and what is carried out first is the row segmentation of nameplate character, to nameplate
Character carries out the projection in y-axis direction, and available nameplate character has one in the pixel distribution in y-axis direction in pixel map
A little places are zero, they correspond to the blank between two rows, and the width of every a line can be calculated according to the pixel distribution in y-axis direction
And the distance between capable and row, divide between the row segmentation for carrying out character picture and then the word for carrying out each line character image
It cuts.It is similar with row segmentation, it is only necessary to the projection of x-axis direction to be done to a certain line character image, available nameplate character is in x-axis side
To pixel distribution, having some places in pixel map is zero, they correspond to two words between blank, can be according to this
The distance of a blank is divided doing word to a line character.For character picture, 32 are compressed it into using the method for linear normalizing
× 32 pixel-matrix, during normalized, if some direction, which first normalizes, reaches 32 dot matrix, another direction is returned
One changes i.e. stopping, can making the destruction that not will cause font during normalized in this way.
In present embodiment, character recognition and correction module identification and arrangement comprising candidate characters, with the small of character
Wave energy and edge orientation histogram construct candidate characters with support vector machine classifier and determine mould as its feature vector
Type carries out optimal alignment to candidate characters by character arrangements model, obtains final nameplate recognition result.Wherein,
The identification of candidate characters carries out wavelet decomposition comprising the character picture after pre-processing, and obtains it in horizontal, vertical, slash
Average energy on these three directions.The average energy of character picture is denoted as Eav, then:Formula
In: f (x, y) indicates the image of some character;F (| | x, y) indicates the absolute value of the image pixel value;M and N indicate the image
It is wide and high.If the low frequency component image obtained after wavelet decomposition is A (f), it contains the profile information of character, high frequency division
Spirogram picture be Bd, j (f), then their average energy be
In formula: d=1,2,3 indicate horizontal, vertical, skim three directions;J=1,2,3 indicate to carry out the number of wavelet decomposition.This
Sample, high fdrequency component image just has 9 energy features, along with the energy feature of low frequency component image, has just obtained a character
10 dimension the feature parameter vectors of image, it may be assumed that
Eav(EavA(f),Eav(B1,1(f)),Eav(B1,2(f),Eav(B3,3(f))),
The edge orientation histogram for extracting character picture needs first to extract the edge shape of character.The side of character is extracted herein
Edge image uses Canny operator.The realization of Canny operator is a multistage treatment process, firstly for image into
Then row Gaussian smoothing converts smoothed out image with Roberts operator, to transformed image, by 360 °
Angular region is divided into 72 grades, and the deflection for calculating normal vector at boundary point in image respectively falls in the frequency in this 72 grades of spaces
Rate has thus obtained the edge orientation histogram vector of character picture.After the feature for extracting character picture, construct herein simultaneously
Training Support Vector Machines classifiers identifies candidate characters.If training character picture is B, Gray-level co-occurrence vector is B=
(B0, B1, B2 ..., B9), edge orientation histogram vector are Y=(Y0, Y1, Y2 ..., Yn), and character picture to be identified is A,
Its Gray-level co-occurrence vector be A=(A0, A1, A2 ..., A9), edge orientation histogram vector be X=(X0, X1, X2 ...,
Xn), T1 and T2 be setting threshold value, then when:
When setting up simultaneously, character B is identified as a candidate characters of character A, and candidate characters may have multiple.In next step
It just needs to do candidate characters correct arrangement, obtains final nameplate recognition result.
The arrangement of candidate characters includes the field term according to used in nameplate, constructs character arrangements model, passes through model pair
Candidate characters are correctly arranged.If the probability for arrangement (m1, m2 ..., mk) the composition term S being made of k candidate characters
For P (S), root
According to N rank Markov model, the probability that term S occurs is only related with the term of front n-1, then:
As N rank Markov character arrangements model.Wherein mI=n+1,...,miAnd mI=n+1,...,mi-1Respectively indicate time
Select character arrangements mI=n+1,...,miAnd mI=n+1,...,mi-1The number occurred in the arrangement of all candidate characters.This k is waited
Word selection symbol finds out it according to above formula and rearranges the probability of different terms by difference, and these probability are arranged from high to low, so
That term of maximum probability will be formed afterwards as final nameplate recognition result.
In present embodiment, what data and the logic check and correction of character recognition and correction module also to malfunction to identification malfunctioned
Data are corrected, and manual correction speed can be improved in this way.
In present embodiment, data processing module, to the facility information that will extract (such as nameplate on Medical Devices
Information, such as title, model etc.) it is matched with the information for being pre-stored in database (such as medical instrument database).And matching is believed
Breath is exported to data memory module.
In present embodiment, data memory module, matched equipment is established electronic record.Preferably, electronics shelves
Case is established according to certain rules, the type for the equipment of such as identification gone out, the date of equipment, etc. of identification gone out.In this way may be used
Planning, such as time-based maintenance, device systems/software upgrading, Fast-Maintenance of failure etc. are managed to equipment according to this data.
Shorten artificial workload in the process.Improve the efficiency of equipment management.
The application also provides a kind of facility information recognition methods, as shown in Fig. 2, this method includes above-mentioned identifying system,
Recognition methods comprises the following steps:
S1 obtains name plate information based on image processing module,
S2, the name plate information that will acquire are pre-processed,
S3, pretreated information carry out character recognition and arrangement and match it with pre-stored information,
Matched information is carried out regulation storage by S4.
In the S2, point for removing inclination and character picture of binaryzation of the pretreatment comprising character picture, character picture
It cuts.It has specifically been described in image pre-processing module, has been not repeated elaboration herein.
The S3, pretreated information carry out character recognition and arrangement and match it with pre-stored information.Its
Using the wavelet energy of character and edge orientation histogram as its feature vector, candidate word is constructed with support vector machine classifier
Decision model is accorded with, optimal alignment is carried out to candidate characters by character arrangements model, obtains final nameplate recognition result.It compares
Traditional OCR identification software, the accuracy rate of identification are improved, and are particularly suited for the identification of equipment nameplate character.
In above embodiment, the image information of nameplate can pass through video camera, smart machine (smart phone) shooting.Shooting
Information afterwards is transmitted to identifying system.
In the design of identifying system, it includes image pre-processing module, character recognition and correction module, data processing
Module, data memory module can be integrated in processing unit.The volume for carrying the identification device of the system can be reduced in this way.Compared with
Good, which connects database by wired or wireless mode.It can conveniently make an inventory information equipment in this way.
It, will be pre- in identified nameplate character information and database in the design of the data processing module of identifying system
The facility information of storage is matched.
In design in the database, according to the information of preset rule memory tool, such as title, model, factory
Family, maintenance information etc..
Through the above description of the embodiments, those skilled in the art can be understood that this specification can borrow
Help software that the mode of required general hardware platform is added to realize.Based on this understanding, the technical solution essence of this specification
On in other words the part that contributes to existing technology can be embodied in the form of software products, the computer software product
Can store in storage medium, such as ROM/RAM, CD, including some instructions are used so that computer equipment (can be with
It is personal computer, server or the network equipment etc.) execute certain parts in each embodiment of this specification or embodiment
The method.
This specification can describe in the general context of computer-executable instructions executed by a computer, such as journey
Sequence module.Generally, program module includes program, component, the data for executing particular task or realizing particular abstract data type
Structure etc..This specification can also be practiced in a distributed computing environment, in these distributed computing environments, by by logical
Communication network and connected remote processing devices execute task.In a distributed computing environment, program module can be located at packet
It includes in the local and remote computer storage media including storing equipment.
The above embodiments merely illustrate the technical concept and features of the present invention, and its object is to allow person skilled in the art
It is to can understand the content of the present invention and implement it accordingly, it is not intended to limit the scope of the present invention.All such as present invention essences
The equivalent transformation or modification that refreshing essence is done, should be covered by the protection scope of the present invention.
Claims (10)
1. a kind of facility information identifying system, which is characterized in that include, image pre-processing module, character recognition and correction module,
Data processing module, data memory module,
Described image preprocessing module receives the information of image collection module transmission and by its binary conversion treatment and will be at binaryzation
Information after reason is transmitted to character recognition and correction module,
The character recognition and correction module receive and identify the information and the corrected information of character recognition are transmitted to number
According to processing module,
The information and pre-stored information progress that the data processing module is transmitted based on the character recognition and correction module
Match and matched information is transmitted to data memory module.
2. the system as claimed in claim 1, which is characterized in that the data processing module connection database is pre- to obtain its
The facility information of storage.
3. the system as claimed in claim 1, which is characterized in that the data memory module is to the character that will identify according to one
It is fixed regularly arranged.
4. the system as claimed in claim 1, which is characterized in that described image preprocessing module also includes the character figure to acquisition
The processing for going inclination, the segmentation of character picture of picture.
5. system as claimed in claim 4, which is characterized in that described image preprocessing module is carried out to the information to equipment
Pretreatment, it includes carry out inclination to character picture to handle.
6. the system as claimed in claim 1, which is characterized in that the pretreatment includes based on global maximum variance threshold value come really
The fixed gray threshold K that binaryzation is carried out to character picture.
7. the system as claimed in claim 1, which is characterized in that the character recognition and correction module are to candidate characters
Identification and arrangement, using the wavelet energy of character and edge orientation histogram as its feature vector, use support vector cassification
Device constructs candidate characters decision model, is arranged candidate characters by character arrangements model, obtains final identification letter
Breath.
8. a kind of facility information recognition methods, it includes systems such as of any of claims 1-7, which is characterized in that
The method comprises the following steps:
S1 obtains name plate information based on image processing module,
S2, the name plate information that will acquire are pre-processed,
S3, pretreated information carry out character recognition and arrangement and match it with pre-stored information,
Matched information is carried out regulation storage by S4.
9. method according to claim 8, which is characterized in that in the S2, binaryzation of the pretreatment comprising character picture, word
Symbol image goes inclination and the segmentation of character picture.
10. method according to claim 8, which is characterized in that in the S2, also comprising tilt to character picture,
The inclination comprising carrying out x-axis to character picture respectively, the projection on y-axis direction, and projected length is respectively L1 and L2, then
The coordinate for the starting point that gray value is 1 in y-axis is detected, the distance to x-axis is denoted as L3, if the tilt angle of going of character picture is
θ, then:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910611848.0A CN110298354A (en) | 2019-07-08 | 2019-07-08 | A kind of facility information identifying system and its recognition methods |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910611848.0A CN110298354A (en) | 2019-07-08 | 2019-07-08 | A kind of facility information identifying system and its recognition methods |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110298354A true CN110298354A (en) | 2019-10-01 |
Family
ID=68030676
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910611848.0A Pending CN110298354A (en) | 2019-07-08 | 2019-07-08 | A kind of facility information identifying system and its recognition methods |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110298354A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111126380A (en) * | 2019-12-02 | 2020-05-08 | 贵州电网有限责任公司 | Method and system for identifying signature of nameplate of power equipment |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101882293A (en) * | 2010-05-19 | 2010-11-10 | 中国农业银行股份有限公司 | Method and system for processing data between domestic custodian bank and overseas custodian agent bank |
CN103473254A (en) * | 2012-06-06 | 2013-12-25 | 三星电子株式会社 | Method and apparatus for storing image data |
CN105335743A (en) * | 2015-10-28 | 2016-02-17 | 重庆邮电大学 | Vehicle license plate recognition method |
CN109784341A (en) * | 2018-12-25 | 2019-05-21 | 华南理工大学 | A kind of medical document recognition methods based on LSTM neural network |
-
2019
- 2019-07-08 CN CN201910611848.0A patent/CN110298354A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101882293A (en) * | 2010-05-19 | 2010-11-10 | 中国农业银行股份有限公司 | Method and system for processing data between domestic custodian bank and overseas custodian agent bank |
CN103473254A (en) * | 2012-06-06 | 2013-12-25 | 三星电子株式会社 | Method and apparatus for storing image data |
CN105335743A (en) * | 2015-10-28 | 2016-02-17 | 重庆邮电大学 | Vehicle license plate recognition method |
CN109784341A (en) * | 2018-12-25 | 2019-05-21 | 华南理工大学 | A kind of medical document recognition methods based on LSTM neural network |
Non-Patent Citations (3)
Title |
---|
何开宇: "《数字化时代的全球个人金融创新》", 31 December 2018 * |
四川省档案局: "《企业档案管理实务》", 30 June 2017 * |
尹远,余正涛: "融合字符及字符排列特征的铭牌识别方法", 《现代电子技术》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111126380A (en) * | 2019-12-02 | 2020-05-08 | 贵州电网有限责任公司 | Method and system for identifying signature of nameplate of power equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10452899B2 (en) | Unsupervised deep representation learning for fine-grained body part recognition | |
Ross et al. | Fingerprint warping using ridge curve correspondences | |
EP3074918B1 (en) | Method and system for face image recognition | |
CN103605972B (en) | Non-restricted environment face verification method based on block depth neural network | |
US7711156B2 (en) | Apparatus and method for generating shape model of object and apparatus and method for automatically searching for feature points of object employing the same | |
CN108985232A (en) | Facial image comparison method, device, computer equipment and storage medium | |
US20120195495A1 (en) | Hierarchical Tree AAM | |
US10803615B2 (en) | Object recognition processing apparatus, object recognition processing method, and program | |
CN104778465B (en) | A kind of matched method for tracking target of distinguished point based | |
CN110490190B (en) | Structured image character recognition method and system | |
CN110717492B (en) | Method for correcting direction of character string in drawing based on joint features | |
CN107392221B (en) | Training method of classification model, and method and device for classifying OCR (optical character recognition) results | |
CN112364881B (en) | Advanced sampling consistency image matching method | |
CN111695609A (en) | Target damage degree determination method, target damage degree determination device, electronic device, and storage medium | |
CN110942473A (en) | Moving target tracking detection method based on characteristic point gridding matching | |
Ouyang et al. | Fingerprint pose estimation based on faster R-CNN | |
CN113033271A (en) | Processing method for learning face identification by using artificial intelligence module | |
CN110298354A (en) | A kind of facility information identifying system and its recognition methods | |
Howe | Inkball models for character localization and out-of-vocabulary word spotting | |
CN109359664A (en) | The efficient Checking model construction method and system of self-teaching update training sample | |
CN117152459A (en) | Image detection method, device, computer readable medium and electronic equipment | |
Tran et al. | 3D point cloud registration based on the vector field representation | |
US20030044067A1 (en) | Apparatus and methods for pattern recognition based on transform aggregation | |
Zhang et al. | Multiview facial landmark localization in RGB-D images via hierarchical regression with binary patterns | |
Wu et al. | A new framework for container code recognition by using segmentation-based and hmm-based approaches |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191001 |
|
RJ01 | Rejection of invention patent application after publication |