CN111352827A - Automatic testing method and device - Google Patents
Automatic testing method and device Download PDFInfo
- Publication number
- CN111352827A CN111352827A CN201811584574.2A CN201811584574A CN111352827A CN 111352827 A CN111352827 A CN 111352827A CN 201811584574 A CN201811584574 A CN 201811584574A CN 111352827 A CN111352827 A CN 111352827A
- Authority
- CN
- China
- Prior art keywords
- verification code
- webpage
- code image
- neural network
- convolutional neural
- 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
- 238000012360 testing method Methods 0.000 title claims abstract description 53
- 238000012795 verification Methods 0.000 claims abstract description 147
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 70
- 238000000034 method Methods 0.000 claims abstract description 35
- 238000012549 training Methods 0.000 claims description 48
- 238000012545 processing Methods 0.000 claims description 20
- 230000015654 memory Effects 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 description 7
- 238000011176 pooling Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 210000002569 neuron Anatomy 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000005336 cracking Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- BUGBHKTXTAQXES-UHFFFAOYSA-N Selenium Chemical compound [Se] BUGBHKTXTAQXES-UHFFFAOYSA-N 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000005520 cutting process 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
- 238000009826 distribution Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000013100 final test Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 229910052711 selenium Inorganic materials 0.000 description 1
- 239000011669 selenium Substances 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3696—Methods or tools to render software testable
-
- 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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention provides an automatic testing method and device, which are used for solving the technical problem of low recognition efficiency when terminal equipment recognizes and cracks a verification code. The method for the automated testing comprises the following steps: acquiring a verification code image of a first webpage; the verification code image comprises at least one verification code character; inputting the verification code image of the first webpage into a convolutional neural network model to obtain the identification content of the verification code image of the first webpage; and loading the identification content into a verification code input box of the first webpage so as to finish the test of the first webpage.
Description
Technical Field
The invention relates to the technical field of computers, in particular to an automatic testing method and device.
Background
In the work of automatic testing, intelligent operation and maintenance and the like, the random graph verification code of the WEB page needs to be automatically identified, and the end-to-end automatic or intelligent operation and maintenance work can be completed. The automatic identification of random pattern verification codes is a common problem in the industry.
Existing solutions generally include two types: firstly, manual input of full-time staff is directly arranged. Secondly, brute force cracking, namely, trying to traverse the whole random graph verification code library by a technical means, dividing regions according to characters, cutting pictures, storing the pictures to the local, and manually marking characters/numbers in the cut graph verification codes; and then comparing whether the picture of the random graph verification code is consistent with the local picture or not during automatic test or intelligent operation and maintenance, if so, outputting the previously marked characters/numbers, and if not, blocking the link.
In the prior art, the verification code is mainly divided and then identified, but most of complex verification code pictures are adhered, so that the division processing is troublesome, and the technical problem of low efficiency in automatic testing in the prior art is caused.
Disclosure of Invention
The embodiment of the invention provides an automatic testing method and device, which are used for solving the technical problem that the efficiency of automatic testing is low in terminal equipment in the prior art.
In a first aspect, an embodiment of the present invention provides an automated testing method, including:
acquiring a verification code image of a first webpage; the verification code image comprises at least one verification code character;
inputting the verification code image of the first webpage into a convolutional neural network model to obtain the identification content of the verification code image of the first webpage;
and loading the identification content into a verification code input box of the first webpage so as to finish the test of the first webpage.
In one possible implementation manner, the convolutional neural network model is obtained by training a sample, where the sample is an authentication code image, and a file name of the sample is an authentication code in the authentication code image;
and taking the verification code image as the input of the convolutional neural network model, verifying the output of the convolutional neural network model and the file name of the sample until the verification code identification accuracy rate reaches a preset threshold value, and determining that the model training is finished.
In one possible implementation, the method further includes:
if the identification content of the verification code image of the first webpage is not obtained, refreshing the webpage, re-obtaining the refreshed verification code image of the second webpage, and inputting the verification code image of the second webpage into the convolutional neural network model.
In one possible implementation, the method further includes:
and taking the verification code image of the first webpage as a sample, and training the convolutional neural network model, wherein the file name of the sample is the verification code in the verification code image of the first webpage.
In a second aspect, an embodiment of the present invention provides an apparatus for automated testing, including:
the acquisition module is used for acquiring a verification code image of a first webpage; the verification code image comprises at least one verification code character;
the processing module is used for inputting the verification code image of the first webpage into a convolutional neural network model to obtain the identification content of the verification code image of the first webpage; and loading the identification content into a verification code input box of the first webpage so as to finish the test of the first webpage.
In one possible implementation manner, the convolutional neural network model is obtained by training a sample, where the sample is an authentication code image, and a file name of the sample is an authentication code in the authentication code image; and taking the verification code image as the input of the convolutional neural network model, verifying the output of the convolutional neural network model and the file name of the sample until the verification code identification accuracy rate reaches a preset threshold value, and determining that the model training is finished.
In one possible implementation, the processing module is further configured to:
if the identification content of the verification code image of the first webpage is not obtained, refreshing the webpage, re-obtaining the refreshed verification code image of the second webpage, and inputting the verification code image of the second webpage into the convolutional neural network model.
In one possible implementation, the processing module is further configured to:
and taking the verification code image of the first webpage as a sample, and training the convolutional neural network model, wherein the file name of the sample is the verification code in the verification code image of the first webpage.
In a third aspect, an embodiment of the present invention provides a computer apparatus, which includes a processor, and the processor is configured to implement the method according to the first aspect when executing a computer program stored in a memory.
In a fourth aspect, the present invention provides a computer-readable storage medium storing computer instructions, which when executed on a computer, cause the computer to perform the method according to the first aspect.
In the embodiment of the invention, the sample image with the mark is input into the convolutional neural network model, and the convolutional neural network model is trained, so that the trained convolutional neural network model can identify the verification code with specific characteristics, and the test efficiency of the automatic test is improved.
Drawings
FIG. 1 is a flow chart of a method of automated testing in an embodiment of the invention;
FIG. 2 is an architecture diagram of a convolutional neural network model provided in an embodiment of the present invention;
FIG. 3 is a block diagram of an apparatus for automated testing in an embodiment of the invention;
FIG. 4 is a block diagram of a computer device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the present invention will be described in further detail with reference to the accompanying drawings, and it is apparent that the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
CAPTCHAs (CAPTCHA) is an abbreviation for "computer Automated reporting test to tellComputers and Humans Apart" (fully Automated Turing test to distinguish computers from Humans), a Public, fully Automated program to distinguish users whether they are computers or Humans, and at present CAPTCHAs with upper and lower case letters and numbers are widely used in large websites to prevent automatic batch registration and repeated login and watering by machines, the letters and numbers themselves being distorted and having much noise and cross-hatching.
With the rapid development and application of internet technology, the network provides abundant resources and great convenience for people, and simultaneously, the security problem of the internet system is accompanied. The appearance of authentication codes is a product of strengthening the security of web systems.
In the automatic test, the equipment is required to automatically identify the verification code, but in the prior art, the end-to-end automatic test and the intelligent operation and maintenance requirements cannot be completely realized by a manual input mode. The brute force cracking method has very low accuracy and very poor flexible adaptability: because the graphic verification code can be automatically and randomly generated according to the algorithm, and the generation strategy of the graphic verification code can be updated in an irregular time, so that the recognition accuracy can be reduced due to slight differences, and particularly, if characters are adhered, partially overlapped or interfered (wavy lines and the like) in a divided area, the intercepted characters or numbers are repeated, invalid or lack a part of characters or numbers, the recognition accuracy can also be reduced, and even the characters or numbers cannot be recognized at all. The manual marking workload is very large, only characters and numbers of verification codes need to be marked, but pictures need to be cut according to regions by using a region segmentation method, so that the marking work is multiplied, continuous tracking and revising are needed, wrong random graphic verification codes are recorded and marked correctly manually, a part of verification codes can be exhausted in a short time, the verification codes can be invalid under the condition of long-term actual use, the searching and comparing efficiency and the identification accuracy are lower and lower along with the increase of marks of the random verification codes, even the random verification codes cannot be used completely, and finally, a large amount of time and manpower are wasted.
Based on the above problem, as shown in fig. 1, an embodiment of the present invention provides an automated testing method, which is applied to an automated testing apparatus, and includes:
step 101: acquiring a verification code image of a first webpage;
wherein the verification code image comprises at least one verification code character;
step 102: inputting the verification code image of the first webpage into a convolutional neural network model to obtain the identification content of the verification code image of the first webpage;
step 103: and loading the identification content into a verification code input box of the first webpage so as to finish the test of the first webpage.
In the embodiment of the invention, the sample image with the mark is input into the convolutional neural network model, and the convolutional neural network model is trained, so that the trained convolutional neural network model can identify the verification code with specific characteristics, and the test efficiency of the automatic test is improved.
A convolutional neural network model used in the embodiment of the present invention is described with reference to the drawings. The convolutional neural network model has a structure as shown in fig. 2. Fig. 2 is a schematic diagram of an architecture of a convolutional neural network model according to an embodiment of the present invention. The convolutional neural network model may be a convolutional module and a fully connected module. Preferably, the neural network comprises 4 layers of convolution modules and a full link layer.
In one possible implementation manner, the convolutional neural network model is obtained by training a sample, wherein the sample is an identifying code image, and the file name of the sample is an identifying code in the identifying code image;
and taking the verification code image as the input of the convolutional neural network model, verifying the output of the convolutional neural network model and the file name of the sample until the verification code identification accuracy rate reaches a preset threshold value, and determining that the model training is finished.
In the embodiment of the invention, when the verification code recognition device can train the convolutional neural network model based on the sample image, the process can be as follows:
first, the verification code recognition device performs preprocessing on a sample image, the preprocessing includes resizing, gray-scale conversion, and the like, and the preprocessed verification code image has a preset pixel size.
The sample image may be a pre-prepared annotated captcha image, and the annotation of the sample image may indicate captcha-related information in the sample image, such as specific captcha characters and/or number of characters, etc. The classifier layer may adjust the output according to the change in the number of characters. The universality of identifying the verification code is increased.
The sample image can be preprocessed, and the preprocessing specifically comprises data cleaning enhancement and invalid feature reduction, graying processing and the like.
Specifically, the preparation of data refreshes the random graph verification code through a selenium module and a chromedriver in python, and snatching and storing according to fixed format size, marking numbers and characters of the random graphic identifying code of the web service system in a manual marking mode, and classifying according to the integral identification of the random graphic identifying code (because the graphic identifying code generally consists of 4 to 6 digits and characters), that is, each category is independent but not exclusive, rather than the way of classifying into single characters, the labeling process is much less than the workload of dividing the same number according to character areas, specifically, the file name of the captured picture is named as the specific corresponding verification code number and character of the random graph verification code (the extension name is not changed), therefore, the training data and the labels are easy to correspond one to one, the processing is convenient, the checking is convenient, and the training and model use efficiency is improved.
In preparation for training, the size of the input verification code image may be adjusted and gradation-converted to adjust the verification code image of the input model to a predetermined pixel size. And for the scene of the graphic verification code without distinguishing colors, converting the multichannel RGB color image into a gray scale image, and further reducing the invalid data characteristics of the image.
For data cleansing enhancement, in particular, the picture sizes may all be uniformly clipped to a uniform size, e.g., 160X 60. And increasing the data thereof by data enhancement means such as picture rotation and the like, and inputting and packaging picture training data according to batches.
By the method, the model trained by using a small amount of effective data (for example, 9000 pictures) and a short training time (for example, 3 hours) can be used, the recognition accuracy of the model trained by using a large amount of data (tens of millions or hundreds of millions of pictures) for a long time (about 1 week) is equivalent, and the training efficiency is improved.
In one possible implementation, the method further includes:
and taking the verification code image of the first webpage as a sample, and training the convolutional neural network model, wherein the file name of the sample is the verification code in the verification code image of the first webpage.
Specifically, the file name corresponds to the label of the training data by labeling the collected data.
By the mode, the training and recognition efficiency is improved, and the automatic testing efficiency is further improved.
Further, the authentication code recognition apparatus inputs the sample images having the marks into the convolutional neural network model in batches according to the number of samples specified in the training, and trains the samples, for example, 80 sample images in each batch. For the training process, please refer to the above description of the architecture of the convolutional neural network model, which is not described herein again.
Furthermore, if the training times are determined to reach the preset times, the identifying code identifying device calculates the average loss rate and the accuracy rate of each batch of sample images, wherein the accuracy rate can be used for representing the probability that the predicted value is matched with the labeled value, and the loss rate is used for representing the number of failed sample identification.
If the accuracy is higher than the set value, the verification code identification device can record the current training parameters and the model of the convolutional neural network model, and the training can be quitted.
In practical application, after the convolutional neural network model is trained, the convolutional neural network model can be tested before being put into use. At this time, the test sample set may be input into the trained convolutional neural network model, the test result of each test sample may be determined, and then each test result may be matched with the label of the corresponding test sample.
If the matching degree is determined to be greater than or equal to the preset matching degree, for example, 80%, the trained convolutional neural network model is high in accuracy and can be used for subsequent identifying of the verification code image. Of course, if the matching degree is lower than the preset matching degree, the model can be continuously trained through the sample image to obtain the more preferable training parameters.
Because the convolutional neural network model in the embodiment of the invention can perform multi-task classification according to the number of characters in a picture to be extracted in an output layer of the network model, and the universality of identifying the identifying code is increased, the number of samples required in training the convolutional neural network model is reduced, for example, under the convolutional neural network model structure in the embodiment of the invention, more than 90% of identifying accuracy can be realized by only 10000 labeled sample patterns, and the accuracy of identifying the image of the identifying code is higher.
After the training parameters and the model are determined, the verification code recognition device can recognize the verification code image through the trained convolutional neural network model. When the verification code image to be recognized is input into the trained convolutional neural network model, the verification code recognition device can preprocess the verification code image to be recognized first so as to adjust the verification code image of the input model to a specified pixel size, and simultaneously can convert the verification code image to be recognized into a gray image.
And then extracting the characteristic information of the verification code image to be identified through the convolution layer and the pooling layer. For example, a convolution kernel with a larger size can be used to extract features in a larger range in the verification code picture, and then a maximized pooling layer with an appropriate size is selected for further feature reduction. And then reducing the size of the volume set core, expanding the number of the volume set core, further performing maximization pooling to extract and simplify characteristic detail characteristics so as to extract more fine characteristic information, and reducing the number of single-channel picture neurons, thereby reducing the weight and the offset of required training and improving the accuracy of the characteristic information.
Furthermore, the number of convolution kernels can be continuously increased through continuous convolution layers, the number of neurons of a single-channel picture is further reduced, and more detailed features are extracted. Subsequently, an anti-overfitting unit which discards neurons with a certain probability can be accessed after the full-connected layer through the full-connected layer in a fully expanded mode. Finally, a certain number of character classifiers are accessed, the classifiers can be adjusted and output according to the change of the number of the characters, namely, the classification of a plurality of tasks can be defined, so that the identification result of the verification code image to be identified, such as the identified verification code, for example, 4 alphabetic characters, and the like, is obtained.
In the embodiment of the invention, before the convolutional neural network model is used for identifying the verification code image, a certain number of sample images are required to be trained.
Then, the sample image with the identification is divided into 3 parts, specifically, a training sample set, a verification sample set and a test sample set, and the size of the batch test set of the training sample set and the verification sample set is specified. Wherein, the training sample set is used for the supervised learning training of the model, and the parameter is adjusted to learn through back propagation; the verification sample set can be used for performing periodic detection on the model in the learning process, and is used for fine-tuning the hyper-parameters of the model so as to optimize the performance of the model; the test sample set can be used for testing the accuracy of the model based on more general data after the model reaches a certain accuracy.
The distribution ratio of the training sample set, the verification sample set and the test sample set can be set as required. In the division, all sample images may be randomly shuffled, for example, the division result may be: the training sample set accounts for 80%, and the validation sample set and the test sample set each account for 10%.
The device may then read each sample set and input the model so that the model identifies the sample image in which the error was noted and process accordingly, e.g., discard.
In addition, the architecture of the convolutional neural network model needs to be set accordingly, and initial values of the weights and the offset parameters need to be set. The processing procedure of the set convolutional neural network model on the verification code image can be as follows:
in the convolution module, a convolution kernel with a larger size can be used for extracting features in a larger range in the verification code picture; the method also can comprise the steps of selecting a maximized pooling layer with a proper size for feature simplification; the number of convolution kernels can be enlarged by using a volume set kernel with a smaller size;
the convolution module in the embodiment of the invention can comprise a pooling layer, and the feature detail feature extraction and simplification are carried out by carrying out maximized pooling so as to extract more small feature information;
furthermore, the convolution module in the embodiment of the invention can continuously perform convolution, and simultaneously continuously expand the number of convolution kernels, further reduce the number of neurons of a single-channel picture, and extract more detailed features;
a full connection layer in the convolutional neural network model can be accessed to an over-fitting prevention unit which discards neurons according to a certain probability; a certain number of character classifiers can be accessed after the full connection layer, and the classifiers can be adjusted and output according to the change of the number of characters, so that the classification of a plurality of tasks can be defined.
After the convolutional neural network model starts to propagate forward and backward, slow decay is performed according to a loss rate, such as a difference between a value calculated by the model forward propagation and a labeled target value, such as a euclidean distance or a cross entropy value. The loss function adopts a mean loss calculation mode of batch samples. During training, a specified number of samples are randomly extracted for training to perform forward propagation calculation, and the average loss rate and the average accuracy rate are calculated according to each batch. After a certain number of training steps, for example 100 times, a verification step is performed, during which the loss rate and the accuracy are calculated, and when the accuracy is higher than a set value, the training parameters and the model of the whole model are maintained, and the process is exited. And the test module acquires the previously trained model parameters and framework, predicts by using the prepared test sample set, and compares the test result with the labeled result to obtain the final test accuracy of the sample batch verification. During actual testing, the model operates on only a single sample image at a time.
In actual use, the pattern verification codes are increased and changed along with the change of time, after the random pattern verification codes with the increased and changed changes are manually labeled, a new convolutional neural network model is retrained, and the identification accuracy is improved along with the increase of training effective data of the convolutional neural network.
After step 103, in one possible implementation, the method further includes:
if the identification content of the verification code image of the first webpage is not obtained, refreshing the webpage, re-obtaining the refreshed verification code image of the second webpage, and inputting the verification code image of the second webpage into the convolutional neural network model.
In the specific implementation process, the model is called for prediction during automatic testing and intelligent operation and maintenance, and the model can be predicted again when failure occurs, which is equivalent to the situation that the error verification code is input again after being manually input, so that the test result in a real scene can be more effectively obtained.
In a specific embodiment, the initially input sample image is 60 × 160 picture data, which is (convolution kernel size, 60,160,1) after random initialization, the output channel is 1, the picture is reduced to (convolution kernel size, 30,80,32) after one time of convolution kernel (1,1) max pooling (2,2), the picture is reduced to (convolution kernel size, 4,10,128) after 3 times of this processing, and a characteristic image of the convolutional neural network which is finally (convolution kernel size, 378) is obtained after one time of full connection (convolution kernel size, 1024). The convolutional neural network integrally identifies and classifies (each class is independent and mutually exclusive) the application algorithm/model of the random pattern verification code, and discriminates and segments the application algorithm/model of the single character number identification and classification. Training for about 5 ten thousand times (about 3 hours) generates a model, and the recognition accuracy of the model on the verification set can reach about 95%.
The identification accuracy of the embodiment of the invention is up to more than 95%, the embodiment of the invention can predict again when failing, the identification accuracy can be infinitely close to 100%, the identification rate of the prior art is basically less than 20%, even the identification can not be carried out at all, and only manual identification can be carried out. The embodiment of the invention is to integrally identify the random graph verification code, firstly, the accuracy and efficiency of identification are accurate and fast compared with the original scheme of character segmentation in actual use comparison, secondly, the manual workload of marking data is very small and can be evaluated, and the original scheme of segmenting according to regional characters can not be identified completely if the characters are adhered, overlapped or interfered (such as wavy lines), and the workload can not be evaluated completely. With the increase of manual label training data, the identification accuracy can be further improved, and with the increase of data, the identification rate of the prior art is inevitably reduced, even cannot be identified at all. The automatic test and intelligent operation and maintenance can be really realized, and the original technology can only be semi-manual and semi-automatic test and operation and maintenance.
As shown in fig. 3, an embodiment of the present invention further provides an identifying device for an authentication code, where the identifying device may be used to implement the method shown in fig. 1, and the method includes:
an obtaining module 301, configured to obtain a verification code image of a first webpage; the verification code image comprises at least one verification code character;
the processing module 302 is configured to input the verification code image of the first webpage into a convolutional neural network model, and obtain identification content of the verification code image of the first webpage; and loading the identification content into a verification code input box of the first webpage so as to finish the test of the first webpage.
In one possible implementation manner, the convolutional neural network model is obtained by training a sample, where the sample is an authentication code image, and a file name of the sample is an authentication code in the authentication code image; and taking the verification code image as the input of the convolutional neural network model, verifying the output of the convolutional neural network model and the file name of the sample until the verification code identification accuracy rate reaches a preset threshold value, and determining that the model training is finished.
In a possible implementation manner, the processing module 302 is further configured to:
if the identification content of the verification code image of the first webpage is not obtained, refreshing the webpage, re-obtaining the refreshed verification code image of the second webpage, and inputting the verification code image of the second webpage into the convolutional neural network model.
In a possible implementation manner, the processing module 302 is further configured to:
and taking the verification code image of the first webpage as a sample, and training the convolutional neural network model, wherein the file name of the sample is the verification code in the verification code image of the first webpage.
As shown in fig. 4, the embodiment of the present invention further provides a computer apparatus, which includes a processor 401 and a memory 402, where the processor 401 is configured to implement the steps of the method for processing a digital certificate update request provided in the first embodiment of the present invention when executing a computer program stored in the memory 402.
Optionally, the processor 401 may specifically be a central processing unit, an Application Specific Integrated Circuit (ASIC), one or more Integrated circuits for controlling program execution, a hardware Circuit developed by using a Field Programmable Gate Array (FPGA), or a baseband processor.
Optionally, processor 401 may include at least one processing core.
Optionally, the electronic device further includes a Memory 402, and the Memory 402 may include a Read Only Memory (ROM), a Random Access Memory (RAM), and a disk Memory. The memory 402 is used for storing data required by the processor 401 in operation. The number of the memories 402 is one or more.
In another embodiment of the present invention, a computer-readable storage medium is further provided, which stores computer instructions, and when the computer instructions are executed on a computer, the computer instructions can implement the steps of the method for processing a digital certificate update request according to an embodiment of the present invention.
In the embodiments of the present invention, it should be understood that the disclosed method and server for processing a digital certificate update request may be implemented in other ways. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, a division of a unit is merely a logical division, and an actual implementation may have another division, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical or other form.
The functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be an independent physical module.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, all or part of the technical solutions of the embodiments of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device, such as a personal computer, a server, or a network device, or a Processor (Processor), to execute all or part of the steps of the methods of the embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a Universal Serial Bus flash drive (USB), a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, and an optical disk.
The above embodiments are only used to describe the technical solutions of the present invention in detail, but the above embodiments are only used to help understanding the method of the embodiments of the present invention, and should not be construed as limiting the embodiments of the present invention. Variations or substitutions that may be readily apparent to one skilled in the art are intended to be included within the scope of the embodiments of the present invention.
Claims (10)
1. A method of automated testing, comprising:
acquiring a verification code image of a first webpage; the verification code image comprises at least one verification code character;
inputting the verification code image of the first webpage into a convolutional neural network model to obtain the identification content of the verification code image of the first webpage;
and loading the identification content into a verification code input box of the first webpage so as to finish the test of the first webpage.
2. The method of claim 1, wherein the convolutional neural network model is obtained by training a sample, wherein the sample is a captcha image, and a file name of the sample is a captcha in the captcha image;
and taking the verification code image as the input of the convolutional neural network model, verifying the output of the convolutional neural network model and the file name of the sample until the verification code identification accuracy rate reaches a preset threshold value, and determining that the model training is finished.
3. The method of claim 1, wherein the method further comprises:
if the identification content of the verification code image of the first webpage is not obtained, refreshing the webpage, re-obtaining the refreshed verification code image of the second webpage, and inputting the verification code image of the second webpage into the convolutional neural network model.
4. The method of claim 3, wherein the method further comprises:
and taking the verification code image of the first webpage as a sample, and training the convolutional neural network model, wherein the file name of the sample is the verification code in the verification code image of the first webpage.
5. An apparatus for automated testing, comprising:
the acquisition module is used for acquiring a verification code image of a first webpage; the verification code image comprises at least one verification code character;
the processing module is used for inputting the verification code image of the first webpage into a convolutional neural network model to obtain the identification content of the verification code image of the first webpage; and loading the identification content into a verification code input box of the first webpage so as to finish the test of the first webpage.
6. The apparatus of claim 5, wherein the convolutional neural network model is obtained by training a sample, wherein the sample is a captcha image, and a file name of the sample is a captcha in the captcha image; and taking the verification code image as the input of the convolutional neural network model, verifying the output of the convolutional neural network model and the file name of the sample until the verification code identification accuracy rate reaches a preset threshold value, and determining that the model training is finished.
7. The apparatus of claim 5, wherein the processing module is further configured to:
if the identification content of the verification code image of the first webpage is not obtained, refreshing the webpage, re-obtaining the refreshed verification code image of the second webpage, and inputting the verification code image of the second webpage into the convolutional neural network model.
8. The apparatus of claim 7, wherein the processing module is further configured to:
and taking the verification code image of the first webpage as a sample, and training the convolutional neural network model, wherein the file name of the sample is the verification code in the verification code image of the first webpage.
9. A computer arrangement, characterized in that the computer arrangement comprises a processor for implementing the method according to any of claims 1-4 when executing a computer program stored in a memory.
10. A computer-readable storage medium having stored thereon computer instructions which, when executed on a computer, cause the computer to perform the method of any one of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811584574.2A CN111352827A (en) | 2018-12-24 | 2018-12-24 | Automatic testing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811584574.2A CN111352827A (en) | 2018-12-24 | 2018-12-24 | Automatic testing method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111352827A true CN111352827A (en) | 2020-06-30 |
Family
ID=71191956
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811584574.2A Pending CN111352827A (en) | 2018-12-24 | 2018-12-24 | Automatic testing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111352827A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105844140A (en) * | 2016-03-21 | 2016-08-10 | 国家电网公司 | Website login brute force crack method and system capable of identifying verification code |
| CN106599900A (en) * | 2015-10-20 | 2017-04-26 | 华中科技大学 | Method and device for recognizing character string in image |
| CN107239778A (en) * | 2017-06-09 | 2017-10-10 | 中国科学技术大学 | The licence plate recognition method of efficiently and accurately |
| CN107368831A (en) * | 2017-07-19 | 2017-11-21 | 中国人民解放军国防科学技术大学 | English words and digit recognition method in a kind of natural scene image |
| CN107454118A (en) * | 2017-10-11 | 2017-12-08 | 北京安博通科技股份有限公司 | Identifying code acquisition methods and device, login method and system |
| CN107967475A (en) * | 2017-11-16 | 2018-04-27 | 广州探迹科技有限公司 | A kind of method for recognizing verification code based on window sliding and convolutional neural networks |
| CN108288078A (en) * | 2017-12-07 | 2018-07-17 | 腾讯科技（深圳）有限公司 | Character identifying method, device and medium in a kind of image |
| CN108563559A (en) * | 2018-03-12 | 2018-09-21 | 平安普惠企业管理有限公司 | A kind of test method of identifying code, device, terminal device and storage medium |
-
2018
- 2018-12-24 CN CN201811584574.2A patent/CN111352827A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106599900A (en) * | 2015-10-20 | 2017-04-26 | 华中科技大学 | Method and device for recognizing character string in image |
| CN105844140A (en) * | 2016-03-21 | 2016-08-10 | 国家电网公司 | Website login brute force crack method and system capable of identifying verification code |
| CN107239778A (en) * | 2017-06-09 | 2017-10-10 | 中国科学技术大学 | The licence plate recognition method of efficiently and accurately |
| CN107368831A (en) * | 2017-07-19 | 2017-11-21 | 中国人民解放军国防科学技术大学 | English words and digit recognition method in a kind of natural scene image |
| CN107454118A (en) * | 2017-10-11 | 2017-12-08 | 北京安博通科技股份有限公司 | Identifying code acquisition methods and device, login method and system |
| CN107967475A (en) * | 2017-11-16 | 2018-04-27 | 广州探迹科技有限公司 | A kind of method for recognizing verification code based on window sliding and convolutional neural networks |
| CN108288078A (en) * | 2017-12-07 | 2018-07-17 | 腾讯科技（深圳）有限公司 | Character identifying method, device and medium in a kind of image |
| CN108563559A (en) * | 2018-03-12 | 2018-09-21 | 平安普惠企业管理有限公司 | A kind of test method of identifying code, device, terminal device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109902622B (en) | Character detection and identification method for boarding check information verification | |
| CN109977943B (en) | Image target recognition method, system and storage medium based on YOLO | |
| US8867828B2 (en) | Text region detection system and method | |
| US10423852B1 (en) | Text image processing using word spacing equalization for ICR system employing artificial neural network | |
| WO2018166116A1 (en) | Car damage recognition method, electronic apparatus and computer-readable storage medium | |
| CN110598686B (en) | Invoice identification method, system, electronic equipment and medium | |
| EP3767536A1 (en) | Latent code for unsupervised domain adaptation | |
| CN106683073B (en) | License plate detection method, camera and server | |
| WO2020164278A1 (en) | Image processing method and device, electronic equipment and readable storage medium | |
| CN111738054B (en) | Behavior anomaly detection method based on space-time self-encoder network and space-time CNN | |
| CN111091080A (en) | Face recognition method and system | |
| CN111813997A (en) | Intrusion analysis method, device, equipment and storage medium | |
| CN110781980A (en) | Training method of target detection model, target detection method and device | |
| CN116311214B (en) | License plate recognition method and device | |
| US20230147685A1 (en) | Generalized anomaly detection | |
| Mazumdar et al. | Universal image manipulation detection using deep siamese convolutional neural network | |
| CN115731422A (en) | Training method, classification method and device of multi-label classification model | |
| CN109101984B (en) | Image identification method and device based on convolutional neural network | |
| CN112818946A (en) | Training of age identification model, age identification method and device and electronic equipment | |
| KR102026280B1 (en) | Method and system for scene text detection using deep learning | |
| US20230069960A1 (en) | Generalized anomaly detection | |
| WO2023173546A1 (en) | Method and apparatus for training text recognition model, and computer device and storage medium | |
| CN111490945A (en) | VPN tunnel flow identification method based on deep learning method and DFI | |
| CN111352827A (en) | Automatic testing method and device | |
| CN111553202B (en) | Training method, detection method and device for neural network for living body detection |
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: 20200630 |