Google not so recently started replacing their reCAPTCHA system with pictures instead of text. Sometimes the images are harder to read than the old text based system, so why not use image recognition to 'bypass' it?
What it does
It processes the images in Google' reCAPTCHA and attempts to programmatically click the right ones.
How I built it
C# code captures the screen and does initial image processing. The data is passed through to a python script that passes the data through to Clarifai. The python script also does the programmatic clicking based on Clarifai's return value(s).
Challenges I ran into
Accomplishments that I'm proud of
It actually works
What I learned
Clarifai is very accurate
What's next for Google reCAPTCHA 'Breaker' - Powered by Clarifai
Implementation of OCR API to read the prompt for full automation.