Deep Learning-Based Phishing Website Detection: Integrating Visual Design Analysis with URL Feature Extraction

  1. Home
  2. December 2025
Sharmin Sultana

Deep Learning-Based Phishing Website Detection: Integrating Visual Design Analysis with URL Feature Extraction

Keywords : Phishing Detection, Deep Learning, Convolutional Neural Network, URL Feature Extraction, Website Security.


Abstract

Phishing attacks remain one of the most prevalent cybersecurity threats, causing significant financial losses and data breaches worldwide. This study presents a comprehensive evaluation of deep learning approaches for phishing website detection by integrating visual design analysis with URL feature extraction. We constructed a dataset comprising 15,000 samples with balanced representation of legitimate and phishing websites, extracting 48 features from URL characteristics and visual design elements. Five deep learning architectures were evaluated: Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), a hybrid CNN-LSTM model, and a Multi-Layer Perceptron (MLP). Experimental results demonstrate that the hybrid CNN-LSTM architecture achieved superior performance with 98.72% accuracy, 98.45% precision, 98.89% recall, and 98.67% F1-score. The findings suggest that combining spatial and sequential feature learning capabilities enhances phishing detection effectiveness, providing a robust framework for real-time web security applications.

Download

Comments
No have any comment !
Leave a Comment