Diatom-Non-neuronal-Cognition

The aim of this project will be to improve upon a Deep Learning model that extracts morphological features from microscopy images of Bacillaria Paradoxa.

Notebooks:

Run in Google Colab - Object detection

Run in Google Colab - Instance segmentation

Run from github

Setup:

• Run sh setup_yolov5.sh for setting up and cloning yolov5 repository. OR
• Run sh setup_scl_yolov4.sh for setting up and cloning Scaled yolov4 repository.

Training:

• Make changes in train_yolov5.sh for training on dataset using yolov5.
  • To start training run sh train_yolov5.sh. OR
• Make changes in train_scl_yolov4.sh for training on dataset using yolov4.
  • To start training run sh train_scl_yolov4.sh.

Prediction:

• Make changes in predict_yolov5.sh for predicting on dataset using yolov5.
  • To make predicitions run sh predict_yolov5.sh. OR
• Make changes in predict_scl_yolov4.sh for predicting on dataset using yolov4.
  • To make predictions run sh predict_scl_yolov4.sh.