Clothing Classifier 🥼

This project is a simple clothing classifier web application built using Streamlit and TensorFlow. It leverages a custom-trained Fashion MNIST model to classify images of clothing items. The application provides real-time predictions along with a bar chart visualizing the confidence levels for each class.

Features

• Upload images of clothing items in JPG, JPEG, or PNG format.
• Classify the uploaded image using a custom-trained Fashion MNIST model.
• Display the predicted clothing category.
• Visualize the confidence levels of all possible classes in a bar chart.

Fashion MNIST Categories

The Fashion MNIST dataset contains 60,000 training images and 10,000 testing images, each belonging to one of the 10 clothing categories. These are:

1. T-shirt/top
2. Trouser
3. Pullover
4. Dress
5. Coat
6. Sandal
7. Shirt
8. Sneaker
9. Bag
10. Ankle boot

These categories represent the different types of clothing that the model has been trained to recognize.

Getting Started

Prerequisites

• Python 3.x
• TensorFlow
• Streamlit
• Pandas
• PIL (Pillow)
• Numpy

Running the App

1. Navigate to the project directory.

   cd fashion_mnist-classifier/app

2. Install dependencies:

   pip install streamlit

3. Start the Streamlit app:

   streamlit run main.py

4. Open the application in your web browser at http://localhost:8501.

Usage

• Upload an image of a clothing item using the file uploader.
• Click the "Classify" button to get the predicted category.
• View the prediction along with a bar chart showing the confidence levels for each class.

File Structure

• main.py: The main Streamlit application file.
• trained_model/: Directory containing the custom-trained model file (trained_fashion_mnist_model.h5).