Hieroglyph Symbol Detection using YOLOv8

This repository provides an end-to-end solution for detecting hieroglyph symbols using YOLOv8, trained from scratch on a custom dataset. The primary goal is to build a lightweight, efficient, and deployable model for hieroglyph detection, with the added flexibility of converting the trained model to TensorFlow.js for web-based applications.

Features

• YOLOv8 Training from Scratch
  • Trained on a hieroglyph dataset with symbols size.
• Custom Dataset Integration
  • Dataset: COTA COCO 50.
• TensorFlow.js Conversion
  • Enables easy deployment in web environments.

Dataset

The dataset used for this project is COTA COCO 50, sourced from Roboflow. It contains annotated hieroglyph images optimized for training and evaluation.

Dataset Overview

• Annotation Format: COCO.
• Classes: Various hieroglyphic symbols.

To access the dataset:

1. Visit COTA COCO 50 on Roboflow.
2. Download the dataset.

Setup and Installation

Prerequisites

• Python 3.8 or above
• PyTorch
• YOLOv8 dependencies

Steps

1. Clone this repository:

   git clone https://github.com/rm-rf-humans/Hieroglyphs.git  
   cd Hieroglyphs  
   

2. Download the dataset from Roboflow and place it in the data/ directory:

data/
├── train/
├── val/
└── test/

Model Conversion to TensorFlow.js

Once training is complete, convert the YOLOv8 model to TensorFlow.js for deployment.

Results

• Evaluation metrics (mAP, precision, recall) are logged after each training epoch.
  • Performance details can be found in the results/ folder.

Contributions

Contributions are welcome! Feel free to open issues or submit pull requests to improve this project.

Acknowledgments

• Roboflow for providing the dataset.
• Ultralytics for the YOLOv8 framework.

For questions or suggestions, please open an issue in the repository.