Symbol Detection on Historical Maps

This folder allows you to train/finetune models for symbol detection on old maps, primarily focusing on Cassini maps. Here, you will find all the processes to prepare and augment your data, as well as train, evaluate, and test your models.

Table of Contents

• Project Structure
• Getting Started
• Installation
• Usage
• Contact

Project Structure

The folder is organized as follows:

├── notebooks/                      # Folder to run notebooks 
│   ├── utils/                      # Utilities functions for the notebooks
│   ├── notebook1.ipynb
│   └── ...
├── models/                         # Folder containing modules for the training of different models
│   ├── detectron2/
│   ├── detr/
│   └── yolo/
├── datasets/                       # Folder containing modules to prepare/augment the datasets
├── data/                           # Folder containing the datasets
├── outputs/                        # Folder with the weights of the different models
├── runs/                           # Folder with training/validation logs
├── tests/                          # Folder for testing different implementations
├───── requirements.txt
└───── README.md

Getting Started

Prerequisites

Make sure you have the following installed:

• Python 3.7 or higher
• Git
• Virtual environment (optional but recommended)

Installation

1. Clone the repository:

   git clone https://github.com/your_username/your_repo.git
   cd your_repo

2. Create a virtual environment:

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

3. Install the dependencies:

   pip install -r requirements.txt

Usage

This part is supposed to be used through the notebooks/ folder. You can know more about the process to follow here.

Implementations

Most of the implementation scripts are used in :

• datasets/ where the modules to manipulate or visualize the datas are stored.
• models/ where the scripts to implement and instantiate the models are stored.
• notebooks/utils utilities functions used by the notebooks.

Datasets

The datasets for training are stored in the datas/ folder and not the datasets/ one ! It is organized as :

models/                        
│   ├── coco_datasets/      #COCO formated datasets
└── └── yolo_datasets/      #YOLO formated datasets

Outputs

The outputs/ directory will contain the weights of the different models after training. You can use these weights for further evaluation or deployment.

Logs

Training and validation logs are stored in the runs/ directory. You can review these logs to monitor the performance and progress of your training sessions.

Contact

Charles Sutty - charles.m.sutty@gmail.com Project Link: https://github.com/your_username/your_repo