Rafeek Sibrikhan

Benchmarking Pre-Trained Feature Extractors: A Comparative Study Across Deep Learning Tasks

This study systematically compares seven pre-trained feature extractors across three architectural families, convolutional neural networks (CNNs), Vision Transformers (ViTs), and self-supervised models to provide practical guidance on model selection for downstream deep learning tasks.