---

EfficientDet for Custom Segmentation

This project involves training the EfficientDet model for segmentation on a custom dataset. The segmentation head utilized in this project is based on the GitHub repository EfficientDetDoesEAST, while the rest of the inferencing and implementation were developed from scratch.

Table of Contents

• Introduction
• Features
• Contributing
• Acknowledgements

Introduction

This project aims to leverage the EfficientDet architecture for segmentation tasks. By integrating the EfficientDet model with a custom segmentation head inspired by the EfficientDetDoesEAST repository, we aim to achieve efficient and accurate segmentation results on our custom dataset.

Features

• EfficientDet-based segmentation
• Custom segmentation head
• End-to-end training and inferencing pipeline
• Support for custom datasets

Inference

The inference pipeline is designed to be user-friendly and efficient. Simply provide the path to your input data and the trained model will generate segmentation results. The output will be saved in the specified directory.

Contributing

We welcome contributions to this project. If you have any suggestions or improvements, please submit a pull request or open an issue.

Acknowledgements

• EfficientDetDoesEAST for the segmentation head implementation.
• The open-source community for their valuable contributions and support.

---