CN113888746A - Welding seam image mode recognition method based on combination of Hu invariant moment and SVM - Google Patents
Welding seam image mode recognition method based on combination of Hu invariant moment and SVM Download PDFInfo
- Publication number
- CN113888746A CN113888746A CN202010630299.4A CN202010630299A CN113888746A CN 113888746 A CN113888746 A CN 113888746A CN 202010630299 A CN202010630299 A CN 202010630299A CN 113888746 A CN113888746 A CN 113888746A
- Authority
- CN
- China
- Prior art keywords
- image
- moment
- welding seam
- svm
- invariant moment
- 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
- 238000003466 welding Methods 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 title claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000012567 pattern recognition method Methods 0.000 claims description 21
- 238000005315 distribution function Methods 0.000 claims description 9
- 239000006185 dispersion Substances 0.000 claims description 6
- 230000005484 gravity Effects 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 238000000605 extraction Methods 0.000 abstract description 7
- 238000012706 support-vector machine Methods 0.000 description 35
- 230000006870 function Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000013519 translation Methods 0.000 description 4
- 230000002035 prolonged effect Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 210000001503 joint Anatomy 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
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
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/245—Classification techniques relating to the decision surface
- G06F18/2451—Classification techniques relating to the decision surface linear, e.g. hyperplane
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a welding seam image mode recognition method based on the combination of Hu invariant moment-SVM, which is used for realizing the recognition of a welding seam image and comprises the following steps: performing feature extraction on the welding seam image based on the Hu invariant moment to obtain seven-dimensional invariant moment data; training an SVM classifier, and determining a hyperplane separating different types of seven-dimensional invariant moment data by solving an optimal value of an objective function; inputting the seven-dimensional invariant moment data into the SVM classifier, and using the hyperplane seven-dimensional invariant moment data separating different types to realize the recognition of the welding seam image mode. The method only retains seven-dimensional data determining the type of the welding seam image, deletes redundant features in a plurality of images, reduces the difficulty of SVM classifier identification, improves the identification time, and still has high identification precision and good robustness even under the condition of low training samples.
Description
Technical Field
The invention relates to the field of data recognition, in particular to a welding seam image mode recognition method based on the combination of Hu invariant moment and SVM.
Background
In order to realize the detection of the welding seam, the system is required to automatically identify various welding seam types. With the development of machine learning, a plurality of technologies capable of realizing image recognition are provided, and common classifiers include Logistic regression, naive Bayes model, k nearest neighbor, Support Vector Machine (SVM), decision tree, random forest and BP neural network. The SVM can be used for linear classification and nonlinear classification. And mapping the data set onto a multidimensional plane by adjusting the kernel function parameters, and converting the two-dimensional linear indivisible problem into the multidimensional linear separable problem.
However, the original welding seam image is identified by directly using the SVM, so that the accuracy is poor, and the identification time is long.
Disclosure of Invention
The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a weld pattern recognition method based on the combination of Hu invariant moment-SVM with high recognition accuracy and short recognition time.
The invention provides a welding seam image mode recognition method based on the combination of Hu invariant moment-SVM, which is used for realizing the recognition of a welding seam image, wherein the welding seam image comprises the following steps: i type welding seam, V type welding seam, T type welding seam, overlap joint type welding seam, circular arc welding seam and no welding seam have such characteristics, include the following step: step 1, performing feature extraction on a welding seam image based on Hu invariant moment to obtain seven-dimensional invariant moment data; step 2, training the SVM classifier, and determining a hyperplane separating different types of seven-dimensional invariant moment data by solving an optimal value of an objective function; and 3, inputting the seven-dimensional invariant moment data into an SVM classifier, and separating different types of seven-dimensional invariant moment data by using a hyperplane so as to realize the identification of the welding seam image mode.
The welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM provided by the invention can also have the following characteristics: the characteristic of Hu invariant moment to weld image extraction is determined by the geometric moment and the central moment of the weld image.
The welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM provided by the invention can also have the following characteristics: wherein the geometric moment m of order (p + q) of the weld imagepqExpressed as:
where f (x, y) is the gray scale distribution function of the successive images.
The welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM provided by the invention can also have the following characteristics: wherein the (p + q) -order central moment mu of the weld imagepqExpressed as:
wherein f (x, y) is a gray scale distribution function of the continuous image,
and
represents the center of gravity of the image and is defined as:
in the formula, m10First moment, m, of the weld image along the x-axis10First moment, m, of the weld image along the y-axis00Is the zeroth order moment of the image.
The welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM provided by the invention can also have the following characteristics: when the welding seam image is a digital discrete image, the step 1 further comprises discretizing the image.
The invention provides a Hu invariant moment-based SVM combinationThe weld image pattern recognition method according to (2), may further have the following features: wherein the (p + q) order geometrical moment m under the dispersionpqExpressed as:
where f (x, y) is the gray scale distribution function of the successive images.
The welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM provided by the invention can also have the following characteristics: wherein the (p + q) order central moment mu in dispersionpqExpressed as:
in the formula, x0Is the central pixel coordinate, y, of the image along the x-axis0Is the coordinate of the central pixel of the image along the direction of the y axis, f (x, y) is the gray distribution function of the continuous image,
for the (p + q) order central moment mu in dispersionpqAfter normalization, the following formula is obtained:
in the formula, npqIs the normalized image (p + q) order central moment, mu00Is the zeroth order center distance of the image.
The welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM provided by the invention can also have the following characteristics: wherein the objective function is
Wherein y is a label of the seven-dimensional invariant moment data, and E ═ E1,e2,...,em) Is m-dimensional normal vector on the hyperplane, and f represents the distance from the hyperplane to the originThe real number of the distance, | E | | | is the norm of the hyperplane, h is the input data sample, and T represents the transpose of the matrix.
The welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM provided by the invention can also have the following characteristics: the pixel size of the weld image is 64 × 64.
Action and Effect of the invention
According to the Hu invariant moment-SVM combination-based weld image pattern recognition method, the image features are extracted by adopting the Hu invariant moment, so that the extracted features are not influenced by image scaling, translation or rotation, the applicability is wider, only seven-dimensional data for determining the type of the weld image are reserved, redundant features in a plurality of images are deleted, the recognition difficulty of an SVM classifier is reduced, and the recognition time is prolonged.
According to the welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM, the SVM classifier is adopted to classify the image data, so that the method still has high recognition accuracy and good robustness even under the condition of low training samples.
Drawings
FIG. 1 is a flow chart of a weld image pattern recognition method based on a Hu invariant moment-SVM combination in an embodiment of the present invention;
FIG. 2 is an image of a no-weld, a V-weld, and an I-weld in an embodiment of the present invention; and
FIG. 3 is an image of a T-weld, lap weld, and arc weld in an embodiment of the present invention.
Detailed Description
In order to make the technical means, the creation features, the achievement purposes and the effects of the invention easy to understand, the invention is specifically described below by combining the embodiment and the attached drawings.
< example >
FIG. 1 is a flow chart of a weld image pattern recognition method based on Hu invariant moment-SVM combination in an embodiment of the present invention.
As shown in fig. 1, the method for recognizing a weld image pattern based on the combination of Hu invariant moment and SVM provided in this embodiment includes the following steps:
step 1, collecting 150 images of six welding seam type images including no welding seam, V type, I type, butt joint, lap joint and circular arc, wherein the total number is 150 multiplied by 6 which is 900 images. Because the pixel size of each welding seam image is 1092 x 964 in the acquisition process, in order to reduce the dimension of the image, the operation speed of a computer is optimized, and the pixel size of each image is modified to 64 x 64 for storage. The modified images are shown in FIGS. 2-3, with each weld image showing only 3 of the images, depending on the context.
Since the shape features of different types of weld images are still different no matter the weld images are shot at any angle in space, in the embodiment, the method for extracting the weld image features is Hu moment invariant. The features extracted by the method are not influenced by image zooming, translation or rotation, and have certain practicability. The Hu invariant moment is mainly determined according to the geometric moment and the central moment of the image to realize the image feature extraction. Setting the gray distribution function of the continuous image to be f (x, y), the (p + q) order geometrical moment of the image is expressed as:
the central moment of (p + q) is expressed as:
in the formula (I), the compound is shown in the specification,
and
represents the center of gravity of the image, which is defined as:
when the image is a digital discrete image, the image needs to be discretized, and the (p + q) order moment under the discretization can be expressed as:
the central moment of (p + q) is expressed as:
normalizing the center moments to obtain:
the Hu invariant moment is a moment which is invariable in zooming, rotation or translation of seven images calculated by second-order and third-order central moments, and the specific expressions of the seven invariable moments are as follows:
by using the above formula, it can be obtained that each image after Hu invariant moment feature extraction contains only seven-dimensional invariant moment data, and the seven-dimensional invariant moment data of each image in fig. 2 is shown in table 1.
TABLE 1
And 2, randomly dividing the seven-dimensional invariant moment data obtained in the step 1 into a training set and a testing set, and importing the training set into an SVM classifier to perform a recognition experiment.
The SVM is a classifier with supervised learning, and can realize two-classification and multi-classification. The main classification idea is to find the ability toA hyperplane separating the different types of data, and requiring the shortest distance of the different types of data that is closest to the hyperplane. Assume that each sample as input data to the SVM classifier is denoted as Pi(x1,x2,…,xm) I-1, 2,3, …, ns (ns is the number of samples), each input sample being an m-dimensional vector. The hyperplane equation separating all the different types of samples is:
ETh+f＝0
wherein E ═ E (E)1,e2,...,em) Is an m-dimensional normal vector on the hyperplane, and f is a real number representing the distance from the hyperplane to the origin. Pi(x1,x2,…,xm) To the hyperplane (E)TAnd, the distance of f) can be expressed as:
in the formula, | E | | | is a norm of the hyperplane, and d is a distance from the sample to the hyperplane. The principle of SVM classifiers is by optimizing the parameters (E) in the hyperplane equationTAnd f) finding the optimal segmentation plane to separate the samples of different types to the maximum extent.
Setting an objective function of SVM optimization as follows:
in the formula, y represents a label of the specimen. Therefore, the optimal value of the objective function is obtained through an optimization algorithm, the optimal hyperplane equation can be obtained, the data of the unknown data type is used as the input of the SVM classifier, and the automatic identification of the unknown data type can be realized through hyperplane separation.
And 3, inputting the seven-dimensional invariant moment data into an SVM classifier, and separating different types of seven-dimensional invariant moment data by using a hyperplane so as to realize the identification of the welding seam image mode.
In this embodiment, 720 samples, 600 samples, 450 samples, and 300 samples are respectively selected from the 900 sample data obtained in step 1 after the Hu invariant moment feature extraction as training sets, the rest are used as test sets, and four groups of training test sets are respectively introduced into the SVM classifier for the recognition experiment and compared with the original acquired image data introduced into the SVM classifier, and the operation results are shown in tables 2 and 3.
TABLE 2
TABLE 3
As shown in tables 2-3, the accuracy of Hu invariant moment-SVM for recognizing different types of weld images is far higher than the accuracy of directly importing the original weld images into an SVM classifier for recognition; and when the number of training samples is reduced, the Hu invariant moment-SVM recognition accuracy is very high, and the robustness is good. In the aspect of recognition time, due to the fact that Hu invariant moment realizes feature extraction of key information in a weld image, only 7-dimensional data for determining the type of the weld image is reserved, compared with the dimension of original data of which the dimension is 64 multiplied by 64 to 4096, redundant features in a plurality of images are deleted, recognition difficulty of an SVM classifier is reduced, and recognition time is prolonged.
Effects and effects of the embodiments
According to the welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM, the Hu invariant moment is adopted to extract the image features, so that the extracted features are not influenced by image scaling, translation or rotation, the applicability is wider, only seven-dimensional data for determining the type of a welding seam image are reserved, a plurality of redundant features in the image are deleted, the difficulty of recognition of an SVM classifier is reduced, and the recognition time is prolonged.
According to the welding seam image pattern recognition method based on the combination of the Hu invariant moment and the SVM, because the SVM classifier is adopted to classify the image data, the method still has high recognition accuracy and good robustness even under the condition of low training samples.
The above embodiments are preferred examples of the present invention, and are not intended to limit the scope of the present invention.
Claims (9)
1. A welding seam image mode recognition method based on Hu invariant moment-SVM combination is used for realizing recognition of welding seam images, and the welding seam images comprise: i type welding seam, V type welding seam, T type welding seam, overlap joint type welding seam, circular arc shape welding seam and no welding seam, its characterized in that includes following step:
step 1, extracting features of the welding seam image based on Hu invariant moment to obtain seven-dimensional invariant moment data;
step 2, training the SVM classifier, and determining a hyperplane separating different types of seven-dimensional invariant moment data by solving an optimal value of an objective function;
and 3, inputting the seven-dimensional invariant moment data into the SVM classifier, and using the hyperplane to separate different types of seven-dimensional invariant moment data so as to realize the recognition of the welding seam image mode.
2. The method for recognizing the weld image pattern based on the combination of Hu invariant moment and SVM according to claim 1, characterized in that:
wherein the characteristics of the Hu invariant moment extracted from the weld image are determined by the geometric moment and the central moment of the weld image.
3. The method for recognizing the weld image pattern based on the combination of Hu invariant moment-SVM according to claim 2, characterized in that:
wherein the geometric moment m of order (p + q) of the weld imagepqExpressed as:
where f (x, y) is the gray scale distribution function of the successive images.
4. The Hu invariant moment-SVM combined weld image pattern recognition method of claim 2, wherein:
wherein the (p + q) -order central moment mu of the weld imagepqExpressed as:
wherein f (x, y) is a gray scale distribution function of the continuous image,
and
represents the center of gravity of the image and is defined as:
in the formula, m10First moment, m, of the weld image along the x-axis10First moment, m, of the weld image along the y-axis00Is the zeroth order moment of the image.
5. The Hu invariant moment-SVM combined weld image pattern recognition method of claim 2, wherein:
wherein, when the weld image is a digital discrete image, step 1 further comprises discretizing the image.
6. The Hu invariant moment-SVM combined weld image pattern recognition method according to claim 5,
wherein the (p + q) order geometrical moment m under the dispersionpqTo representComprises the following steps:
where f (x, y) is the gray scale distribution function of the successive images.
7. The Hu invariant moment-SVM combined weld image pattern recognition method according to claim 5,
wherein the (p + q) order central moment mu in dispersionpqExpressed as:
in the formula, x0Is the central pixel coordinate, y, of the image along the x-axis0Is the coordinate of the central pixel of the image along the direction of the y axis, f (x, y) is the gray distribution function of the continuous image,
for the (p + q) order central moment mu at said dispersionpqAfter normalization, the following formula is obtained:
in the formula, npqIs the normalized image (p + q) order central moment, mu00Is the zeroth order center distance of the image.
8. The Hu invariant moment-SVM combined weld image pattern recognition method according to claim 1,
wherein the objective function is
Wherein y is a label of the seven-dimensional invariant moment data, and E ═ E1,e2,...,em) Is an m-dimensional normal vector on the hyperplane, f represents hyperplaneThe real number of the face-to-origin distance, | E | | | is the norm of the hyperplane, h is the input data sample, and T represents the transpose of the matrix.
9. The Hu invariant moment-SVM combined weld image pattern recognition method according to claim 1,
wherein the pixel size of the weld image is 64 x 64.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010630299.4A CN113888746A (en) | 2020-07-01 | 2020-07-01 | Welding seam image mode recognition method based on combination of Hu invariant moment and SVM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010630299.4A CN113888746A (en) | 2020-07-01 | 2020-07-01 | Welding seam image mode recognition method based on combination of Hu invariant moment and SVM |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113888746A true CN113888746A (en) | 2022-01-04 |
Family
ID=79013016
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010630299.4A Pending CN113888746A (en) | 2020-07-01 | 2020-07-01 | Welding seam image mode recognition method based on combination of Hu invariant moment and SVM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113888746A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115880288A (en) * | 2023-02-21 | 2023-03-31 | 深圳市兆兴博拓科技股份有限公司 | Detection method and system for electronic component welding and computer equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104463122A (en) * | 2014-12-09 | 2015-03-25 | 四川大学 | Seal recognition method based on PCNN |
| CN106022375A (en) * | 2016-05-19 | 2016-10-12 | 东华大学 | HU invariant moment and support vector machine-based garment style identification method |
| CN107194383A (en) * | 2017-07-10 | 2017-09-22 | 上海应用技术大学 | Based on improving Hu not bending moment and ELM traffic mark board recognition methods and device |
| CN108178037A (en) * | 2017-12-30 | 2018-06-19 | 武汉大学 | A kind of elevator faults recognition methods based on convolutional neural networks |
| CN109115812A (en) * | 2018-08-23 | 2019-01-01 | 中国石油大学(北京) | A kind of weld seam egative film defect identification method and system |
| CN112036435A (en) * | 2020-07-22 | 2020-12-04 | 温州大学 | Brushless direct current motor sensor fault detection method based on convolutional neural network |
-
2020
- 2020-07-01 CN CN202010630299.4A patent/CN113888746A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104463122A (en) * | 2014-12-09 | 2015-03-25 | 四川大学 | Seal recognition method based on PCNN |
| CN106022375A (en) * | 2016-05-19 | 2016-10-12 | 东华大学 | HU invariant moment and support vector machine-based garment style identification method |
| CN107194383A (en) * | 2017-07-10 | 2017-09-22 | 上海应用技术大学 | Based on improving Hu not bending moment and ELM traffic mark board recognition methods and device |
| CN108178037A (en) * | 2017-12-30 | 2018-06-19 | 武汉大学 | A kind of elevator faults recognition methods based on convolutional neural networks |
| CN109115812A (en) * | 2018-08-23 | 2019-01-01 | 中国石油大学(北京) | A kind of weld seam egative film defect identification method and system |
| CN112036435A (en) * | 2020-07-22 | 2020-12-04 | 温州大学 | Brushless direct current motor sensor fault detection method based on convolutional neural network |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115880288A (en) * | 2023-02-21 | 2023-03-31 | 深圳市兆兴博拓科技股份有限公司 | Detection method and system for electronic component welding and computer equipment |
| CN115880288B (en) * | 2023-02-21 | 2023-11-14 | 深圳市兆兴博拓科技股份有限公司 | Detection method, system and computer equipment for electronic element welding |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Afzal et al. | Cutting the error by half: Investigation of very deep cnn and advanced training strategies for document image classification | |
| Wang et al. | Feature context for image classification and object detection | |
| Prashanth et al. | Classification of handwritten Devanagari number–an analysis of pattern recognition tool using neural network and CNN | |
| Tarawneh et al. | Invoice classification using deep features and machine learning techniques | |
| US8488883B2 (en) | Robust and efficient image identification | |
| Zhai et al. | OCR-based neural network for ANPR | |
| Faraki et al. | Fisher tensors for classifying human epithelial cells | |
| Cilia et al. | An end-to-end deep learning system for medieval writer identification | |
| Bhowmik et al. | Handwritten Bangla word recognition using HOG descriptor | |
| Chen et al. | Structured document classification by matching local salient features | |
| Neena et al. | Image classification using an ensemble-based deep CNN | |
| Choudhury et al. | Handwritten bengali numeral recognition using hog based feature extraction algorithm | |
| Deore et al. | Histogram of Oriented Gradients Based Off-Line Handwritten Devanagari Characters Recognition Using SVM, K-NN and NN Classifiers. | |
| Shivakumara et al. | Fractional means based method for multi-oriented keyword spotting in video/scene/license plate images | |
| Nair et al. | An Enhanced Approach for Binarizing and Segmenting Degraded Ayurvedic Medical Prescription. | |
| Hassan et al. | Arabic handwriting word recognition based on scale invariant feature transform and support vector machine | |
| Vyas et al. | Classification of handwritten Gujarati numerals | |
| Sah et al. | Text and non-text recognition using modified HOG descriptor | |
| Susanto et al. | Histogram of gradient in k-nearest neighbor for Javanese alphabet classification | |
| Bakheet et al. | Content-based image retrieval using brisk and surf as bag-of-visual-words for naïve Bayes classifier | |
| Hassan et al. | Arabic handwriting text recognition based on efficient segmentation, DCT and HOG features | |
| CN113888746A (en) | Welding seam image mode recognition method based on combination of Hu invariant moment and SVM | |
| Ali et al. | Different handwritten character recognition methods: a review | |
| Jubair et al. | A simplified method for handwritten character recognition from document image | |
| Allaf et al. | Automatic recognition of artistic arabic calligraphy types |
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 |