🍓 Auto-Strawberry: Strawberry Daughter Plant Dataset and Pretrained Models

Welcome to the Auto-Strawberry project! This repository hosts the strawberry daughter plant dataset and pretrained deep learning models developed for automated estimation of plant growth attributes using stereo vision and deep learning techniques.

📦 Dataset Overview

The Strawberry Daughter Plant Dataset contains:

• Multi-view RGB images: 6 views per sample captured using a stereo vision camera setup (cam0 and cam1).
• Ground truth labels:
  • Total Leaf Area
  • Fresh Mass
  • Largest Petiole Length
  • Average Crown Diameter
• Format: Images in .jpg and labels stored in a JSON file.

Dataset Structure:

- cam0/
- cam1/
- annotations.json

🧠 Pretrained Models

Available Backbone Models:

• ResNet34
• Vision Transformer (ViT-B-16)
• EfficientNet B0

Model Weights:

Each model was trained using the strawberry dataset for the regression task of estimating leaf area and other growth metrics. The weights provided here can be used for inference and fine-tuning.

📥 Download and Usage

Hugging Face Integration:

pip install huggingface_hub

from huggingface_hub import hf_hub_download

# Download the model from Hugging Face Hub
hf_hub_download(repo_id="sinabjam/auto-strawberry", filename="desired_file_name.pth")

🎯 Applications

• Plant phenotyping
• Automated growth monitoring
• Precision agriculture insights

📜 License

This project is licensed under the Apache 2.0 License.

📬 Contact

• Author: Sina Baghbanijam
• Email: sbaghba@ncsu.edu
• Project Repository: GitHub Link

🌟 If you use this dataset or models in your research, please cite this repository!