Text-based CAPTCHA strengths and weaknesses

By Elie Bursztein, Matthieu Martin, John C. Mitchell. Conference on Computer and Communications Security, 2011

374 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Inference Compilation and Universal Probabilistic Programming (2016, methodology): “Faceook Captchas are collected from a page Initially obtaining low recognition rates (< 10%), with several iterations of model modifications (involving tuning of the prior distributions for font size and renderer noise), we were able to achieve 81% and 42% recognition rates with real Wikipedia and F”
The End is Nigh: Generic Solving of Text-based CAPTCHAs (2014, methodology): “[10].”
A generative vision model that trains with high data efficiency and breaks text-based CAPTCHAs (2017, methodology): “A CAPTCHA is considered broken if it can be automatically solved at a rate above 1% (3).”
I am Robot: (Deep) Learning to Break Semantic Image CAPTCHAs (2016, background): “[35] conducted an extensive study on the strengths and weaknesses of text captchas.”
Fake Co-visitation Injection Attacks to Recommender Systems (2017, background): “We note that recent studies [28, 29] demonstrated that CAPTCHA challenges can be automatically solved by machine learning techniques with relatively high accuracies, and an attacker can outsource CAPTCHA challenges to human workers using crowdsourcing platforms [30].”