Gemstone Classifier

Authors Stephan Amann, Tanja Huber, David Kleindiek

Date 15.03.2024

GitHub repository for the final project of the 'Machine Learning and Image Processing ' class during the winter term at the Eberhard-Karls Universität Tübingen.

The project report is available here

Overview

This repository contains the research work conducted to develop a gemstone classifier using machine learning techniques. The aim of this project is to accurately classify different types of gemstones based on their visual features, aiding in the identification process.

Objectives

• Scrape gemstone images from online sources to create a diverse and representative dataset.
• Preprocess the collected data, including resizing images, handling missing values, and encoding categorical variables, to prepare it for training a machine learning model.
• Train a machine learning model on the prepared dataset and evaluate its performance using appropriate metrics to assess accuracy and reliability.

Repository Structure

The repository is organized into several directories:

• src/: Contains Python scripts for data scraping, preprocessing, model training, and evaluation.
• mod/: Contains trained model files, including the results of each epoch during training for each dataset.
• dat/: Contains raw and processed data files used in the project, including images of gemstones, metadata, and intermediate datasets.
• doc/: Documentation, including the project report and any additional figures.

Users need to obtain the model files and datasets themselves due to their large file sizes.

Running the code

1. Clone the repository:

git clone https://github.com/SATHDKTT/bv-ml-project.git
cd bv-ml-project

2. Set Up a Virtual Environment (Optional but Recommended):

python -m venv venv
source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install Required Packages:

pip install -r requirements.txt

4. Run The Notebooks:

jupyter notebook src/