Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A slider notch positioning and dragging trajectory generation method for slider verification code

A trajectory generation and verification code technology, applied in the Internet field, can solve problems such as insufficient randomness, repeated trajectory generation, and a cliff-like decline in the recognition pass rate, and achieve the effect of strong pass rate stability and high recognition pass rate

Active Publication Date: 2021-05-04
南京星云数字技术有限公司
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The above three trajectory generation methods are mainly based on the fitting of a fixed mathematical model and the retrieval of trajectory databases. The algorithm itself has not really learned the essential characteristics of human behavior of dragging the slider, so there are defects of insufficient randomness and possible repetition of generated trajectories. , especially when the slider verification party uses artificial intelligence and big data technology, the above three methods will have a cliff-like decline in the recognition pass rate as the number of tests increases

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A slider notch positioning and dragging trajectory generation method for slider verification code
  • A slider notch positioning and dragging trajectory generation method for slider verification code
  • A slider notch positioning and dragging trajectory generation method for slider verification code

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach 1

[0222] combine image 3 , the scheme first uses the Burr statistical model to model the relationship between the displacement distance and the dragging time, that is, the model outputs the total time-consuming A that conforms to the Burr model, and the unit is milliseconds.

[0223] Secondly, according to the collected human slider drag data set, the statistical relationship of the time sampling frequency is analyzed, and the model is established to generate the time sampling sequence. In this step, the acceptance-rejection method (Acceptance-RejectionMethod) is first used to select the non-uniform sampling points of the time series, and then for the sampling points that have been determined to be selected, the specific sampling values ​​(in milliseconds) of the sampling points are calculated. Gaussian random number generation with specified mean and variance. This step will finally generate a non-uniformly sampled time series (the total time consumption of the series is A) a...

Embodiment approach 2

[0225] This implementation scheme uses a layer of LSTM to directly establish a model from input (displacement distance, time-consuming time) to output non-uniform sampling time series. All the time step output of the LSTM layer will be used as the input of the LSTM backend, and the number of time steps is 55, and the number of time steps is filled with zeros.

[0226] combine image 3 , the LSTM regression model is written using the Tensorflow framework. Its network architecture and loss function have been customized and modified to adapt to the generation of variable-length slider drag trajectories with different displacement distances and different drag times in the application of slider verification code recognition. First, the input of the LSTM model is forced into the actual time series length, and the output is the full time step output; second, a dropout layer is added before the output layer is connected. The dropout layer and the output layer use fully connected (ful...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a method for positioning the gap of a slider and generating a dragging track of a slider verification code. The method includes two parts, offline model training: building a target detection model based on deep learning, calculating the coordinates of the center position of the slider notch according to the background image of the slider notch, and marking a rectangular label box for supervised target detection model training; The human slider drags the trajectory, establishes the LSTM front-end model and the LSTM back-end regression model, and performs model training. Online model service: Take a screenshot of the screen, process it to obtain the slider picture and the background picture of the slider gap; take the background picture of the slider gap as input, execute the trained target detection model, and obtain the coordinates of the center position of the slider gap; execute The trained LSTM front-end model and the LSTM back-end regression model obtain the time series of the slider drag trajectory. The invention realizes the random generation of the trajectory of the anthropomorphic slider, and has high recognition pass rate and strong pass rate stability.

Description

technical field [0001] The invention relates to the technical field of the Internet, in particular to a method for locating a slider gap and generating a drag trajectory of a slider verification code. Background technique [0002] Captcha is a kind of Turing test for determining whether an operation is performed on a human. In the field of software testing, testers use software to conduct automated testing of websites, rapid release of information, and automated transactions. Automatic identification of verification codes can help improve testing efficiency and test coverage, reduce manual testing links, and avoid human errors. The slider verification code is a specific implementation of the Turing test, and its implementation type is called Dynamic CognitiveGame. At present, the industry is researching a complete end-to-end slider verification code automatic identification technology, which includes research on the system architecture level and algorithm level. At the alg...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F21/36
CPCG06F3/04847G06F21/36G06N3/084G06N3/045G06N3/044
Inventor 王元
Owner 南京星云数字技术有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products