Scholarlink

CAPTCHAs (Completely Automated Public Turing tests to tell Computers and Humans Apart) have long been used to prevent automated bots from accessing online services. However, advancements in deep learning have enabled AI-based models to bypass traditional CAPTCHA systems. This paper presents a comparative analysis of deep learning models, primarily based on Convolutional Neural Networks (CNNs), in the context of CAPTCHA prediction. The models are evaluated based on accuracy, efficiency, and robustness. Additionally, security and ethical concerns regarding AI-driven CAPTCHA breaking are explored. The findings highlight the necessity of innovative CAPTCHA mechanisms to maintain secure authentication.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep Learning Models for CAPTCHA Prediction: A Comparative Analysis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Devarshi Dave,
  • Dheeraj Kumar Shringi,
  • Nirav Bhatt

摘要

CAPTCHAs (Completely Automated Public Turing tests to tell Computers and Humans Apart) have long been used to prevent automated bots from accessing online services. However, advancements in deep learning have enabled AI-based models to bypass traditional CAPTCHA systems. This paper presents a comparative analysis of deep learning models, primarily based on Convolutional Neural Networks (CNNs), in the context of CAPTCHA prediction. The models are evaluated based on accuracy, efficiency, and robustness. Additionally, security and ethical concerns regarding AI-driven CAPTCHA breaking are explored. The findings highlight the necessity of innovative CAPTCHA mechanisms to maintain secure authentication.