Abstract
Nowadays, almost all major websites employ CAPTCHAs. This prevents website scraping, fake account creation as well as DDoS or bruteforce attacks. For anonymity reasons, mainstream CAPTCHAs such as Google’s reCAPTCHA cannot be used on the darkweb. Due to the evolution of machine learning and computer vision, the CAPTCHA challenges used there, such as the clock CAPTCHA, are usually more arduous than those found on the clearweb. This paper presents an automated system that uses machine learning to break clock CAPTCHA challenges with a high success rate. We evaluate our system in a real world setting against 725 clock challenges from live darkweb marketplaces. Our results show an accuracy of 96.83% while maintaining low time requirements while analyzing, predicting and submitting the CAPTCHA solution.
Original language | English |
---|---|
Title of host publication | Proceedings of the 19th International Conference on Security and Cryptography |
Editors | Sabrina De Capitani di Vimercati , Pierangela Samarati |
Volume | 1 |
Publisher | SCITEPRESS Digital Library |
Publication date | 2022 |
Pages | 357-365 |
ISBN (Electronic) | 978-989-758-590-6 |
DOIs | |
Publication status | Published - 2022 |
Event | 19th International Conference on Security and Cryptography (SECRYPT 2022) - Lisbon, Portugal Duration: 11 Jul 2022 → 13 Jul 2022 Conference number: 19 https://secrypt.scitevents.org/ |
Conference
Conference | 19th International Conference on Security and Cryptography (SECRYPT 2022) |
---|---|
Number | 19 |
Country/Territory | Portugal |
City | Lisbon |
Period | 11/07/2022 → 13/07/2022 |
Internet address |
Series | International Conference on Security and Cryptography - SECRYPT - Proceedings |
---|---|
ISSN | 2184-7711 |