By MISHA DEY
A Text-Based CAPTCHA(Completely Automated Public Turing Test) Breaking System in Python Using CNN(Convolutional Neural Networks). Dataset: Generated using the "Captcha" library.
The tern 'CAPTCHA' Stands for Completely Automated Public Turing Test.
Libraries/Modules Used :
Dataset Description :
The train and test dataset was generated using the "captcha" module/library, used to generate captchas/verification codes for the given set of characters.
1. Captcha Sequence Generator: It is used for the real-time data feeding to the CNN model
Captcha Character Set = Set of all Uppercase Alphabets + Set of all Lowercase Applabets + Set of digits (0-9)
" __init__ " function used for initialization
"__len__" return the number of batches per epochs
"__getitem__" It is executed when a batch corresponding to a batch index is called to generate it.
2. Convolutional Neural Network Model :
Total params: 2,993,494 Trainable params: 2,990,550 Non-trainable params: 2,944
Overall Accuracy = 0.95578125
DEMO TEST RESULT :
Demonstration Video Link :