Image Annotation
This project focuses on the creation of a high-quality polygon-annotated image dataset of website screenshots to support visual phishing detection using deep learning. Traditional phishing detection methods rely heavily on URL analysis and textual features, but often fail against visually convincing phishing pages. To address this limitation, this project applies polygon-based image annotation to accurately label irregular and overlapping webpage elements such as login forms, buttons, logos, text fields, and URL bars.