Fashion-MNIST Image Classification: CNN vs. GAN-Enhanced CNN

This project focused on comparing the performance of a traditional Convolutional Neural Network (CNN) against a Generative Adversarial Network (GAN)-Enhanced CNN for classifying images in the Fashion-MNIST dataset. The goal was to explore how GAN-generated synthetic data could improve the accuracy of the CNN model for better image classification.

Impact: This project demonstrated the potential of GANs in enhancing model performance by generating synthetic data, a crucial factor when the dataset is small or imbalanced.

Challenges and Achievements: Implementing the GAN and integrating it with the CNN posed some complexities. However, the final model showed improved performance, proving the value of synthetic data in model training.