This repository showcases a comprehensive approach to fashion image classification using the Maryland Fashion Dataset. It includes four scenarios: building a CNN from scratch, leveraging transfer learning with VGG16, fine-tuning the VGG19 model, and implementing an autoencoder for image representation. Each scenario demonstrates various techniques.
computer-vision deep-learning tensorflow keras autoencoder image-classification convolutional-neural-networks transfer-learning vgg16 vgg19 data-augmentation fine-tuning fashion-classification cleaned-maryland maryland-polyvore
-
Updated
Oct 6, 2024 - Jupyter Notebook