Skip to content

DevoLearn/Model-Annotation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Model Annotation for Cell Tracking Challenge dataset

In conjunction with the OpenWorm Foundation and Orthogonal Research and Education Lab

C. elegans Image Annotation for Hacktoberfest

Hello, contributors! 🎃 Welcome to our Hacktoberfest project. We're aiming to annotate a dataset of C. elegans images to fine-tune the Detection 2 model and SAM segmentation for an upcoming GSoC project. Your contribution will play a pivotal role in advancing research in this area.

📂 Dataset Structure

  • images/: Contains the C. elegans images.
  • segmentation_maps/: Contains the segmentation maps for the images.
  • data.csv: A CSV file listing the image file names.

🎯 Task

We need bounding boxes around each C. elegans cell in the images. dataset link, csv_file link

Guidelines:

  1. Bounding Box: The bounding box should tightly enclose the cell without cutting any part of it.
  2. Overlapping Cells: In cases where cells overlap, draw individual bounding boxes for each visible cell.
  3. Avoid Background: Ensure that the bounding box only contains the cell and as little background as possible.

🛠 Tools

For this project, we'll be using the open-source tool MakeSense.

Steps to use MakeSense:

  1. Visit MakeSense.
  2. Click on Upload Images and select the C. elegans images you want to annotate.
  3. Define a label, e.g., "C. elegans cell".
  4. Start drawing bounding boxes following the guidelines provided.
  5. Once done, click on the Export button and select the COCO (JSON) format.
  6. Save the exported JSON annotations locally.

📝 How to Contribute

  1. Fork this repository.
  2. Clone your forked repository to your local machine.
  3. Create an Issue: Before starting the annotation, create an issue in the main repository indicating that you're taking up 10 images for annotation. Mention the exact image names in the order they appear in data.csv to ensure clarity and avoid overlapping work.
  4. Pick the next 10 images in order from the data.csv that haven't been annotated yet.
  5. Annotate the images using MakeSense and the guidelines provided.
  6. Save the annotations in COCO JSON format.
  7. Create a new folder with your name (e.g., JohnDoe/). Inside this folder, create two subfolders: images/ and annotations/.
  8. Place the annotated images in the images/ folder and the COCO JSON annotations in the annotations/ folder.
  9. Update the data.csv file by adding your name next to the images you've annotated.
  10. Commit your changes with a meaningful commit message.
  11. Push your changes to your forked repository.
  12. Create a Pull Request to the main repository with a title like "Annotations by [Your Name]". In the PR description, mention Close #[Issue Number] to link and close the issue you created.
  13. Once your PR is reviewed and merged, you can pick another set of images to annotate if you wish!

🎁 Rewards

  • Every valid PR will count towards your Hacktoberfest contributions.
  • Top contributors will be acknowledged in the project's documentation and chance for the upcoming gsoc project.

🤝 Support

If you have any questions or face any issues, please open an issue in this repository, and we'll get back to you as soon as possible.

Happy annotating and happy Hacktoberfest! 🍂🎉

How to participate:

Watch this video on Detectron 2 link

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published