"PhishVisNet: A Multimodal Visual Structural Dataset for Advanced Phishing Website Detection"

"This dataset provides a comprehensive collection of visual and structural web assets for training and benchmarking advanced phishing detection models, including Quantum Convolutional Neural Networks (QCNNs). Unlike traditional URL-based datasets, it focuses on visual design, interface structure, logo integrity, and webpage patterns across banking, government, financial, authentication, and general web platforms. It includes both legitimate and phishing instances collected from diverse real-world scenarios to enable robust detection of deceptive websites.The dataset contains over 4,428 labeled samples, including webpage screenshots, logos, HTML files, and DOM graph representations. It features balanced legitimate and phishing samples covering multiple attack strategies. Each sample captures visual and structural cues such as layout inconsistencies, branding misuse, authentication manipulation, and code-level similarities. The dataset supports image-based, graph-based, and multi-modal cybersecurity research and benchmarking."