Skip to content

jayanth0075/Detecting-Bacterial-Flagellar-motors--Using-YOLOV8L

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 

Repository files navigation

Detecting bacterial flagellar motors in electron microscopy images using YOLOv8 object detection.

Why detecting bacterial motors !?

Manually identifying bacterial flagellar motors in microscopy images is time-consuming and requires expert knowledge. This automated detection system can:

  • Speed up the analysis workflow for biologists
  • Provide consistent detection across different images
  • Help researchers process large datasets efficiently
  • Serve as a foundation for more advanced detection systems

Results

  • The map50 score was 87.75
  • the map 50-95 score was 51.9

The model performs well but has room for improvement for higher IOU thresholds - perfect for further research and optimization!

What I Used

Model & Framework

  • YOLOv8l (Large variant)
  • Ultralytics 8.3.176
  • PyTorch 2.6.0+cu124

Hardware

  • GPU: NVIDIA Tesla T4 (15GB VRAM) [ Use collab pro for better perfomance using good gpu with more RAM]
  • Platform: Google Colab

Training Configuration

  • Epochs: 200(put patience level )
  • Image Size: 940px
  • Batch Size: 6
  • Patience: 10 (early stopping)

Data Augmentation

  • Rotation (±10°)
  • Translation (10%)
  • Scaling (20%)
  • Horizontal Flip (50%)
  • Mosaic Augmentation

The complete code is available in this repository. Just mount your Google Drive, upload the dataset, and you're good to go!

Research Collaborations & Contributions

I'm Jayanth , I am actively working on multiple research projects and am open to collaborations in machine learning, computer vision, and biological image analysis!

If you're interested in:

  • Improving this bacterial motor detection system
  • Exploring novel architectures for microscopy images
  • Joint research projects in ML/CV
  • Publishing research papers together
  • Any other research collaboration

Please feel free to reach out !!

Email: jayanth9b.vhs@gmail.com or jayanthadavi@gmail.com

I'm looking for people who want to collaborate, contribute, and build something meaningful together. Whether you're a researcher, student, or just someone passionate about ML - let's connect and workhard !

About

Locating bacterial flagellar motors in cryo-electron tomography slices using YOLOv8l. Fine-tuned on the BYU Kaggle 2025 dataset with mosaic and geometric augmentations, achieving 87.5% mAP50 and 0.52 mAP50-95 on a single T4 GPU.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors