Clothing Classification with Convolutional Neural Networks (CNNs) using Fashion-MNIST

This project implements a Convolutional Neural Network (CNN) using TensorFlow and Keras to classify clothing images from the Fashion-MNIST dataset.

Key Highlights: Preprocessing: Images were normalized (scaled between 0 and 1) and labels were transformed using one-hot encoding. Architecture: The model features Conv2D layers for feature extraction, MaxPooling2D for downsampling, and Dense layers for final classification, incorporating Dropout to prevent overfitting. Results: The model achieved an accuracy of approximately 91% on the test set. Evaluation: Learning curves, a confusion matrix to detect misclassified categories, and a visual comparison of correct vs. incorrect predictions were generated to assess performance.