The article discusses the development of SEShield, a framework designed to detect social engineering (SE) attacks in-browser. It consists of three components: SECrawler, a custom security crawler that gathers examples of SE attacks; SENet, a deep learning-based image classifier trained on SECrawler data that detects traits of SE attack pages; and SEGuard, an extension that incorporates SENet into the web browser for real-time SE attack detection. The system has shown to detect new SE attacks with a detection rate of up to 99.6% at 1% false positive. This presents an effective first defense against SE attacks on the web.
Publication date: 15 Jan 2024
Project Page: Not provided
Paper: https://arxiv.org/pdf/2401.05569
