An Interactive Deep Learning-Based Segmentation Platform for 3D Datasets

PHILOW, a human-in-the-loop deep learning workflow, was developed to enable segmentation through deep learning. It’s characterized by its ability to perform all processes — from annotation to model training, inference, retraining, and final correction. However, since PHILOW initially focused on isotropic 3D images, it does not support a wide variety of images and lacks sufficient documentation and demonstrations for biologists.

This project seeks to support more image types, documentation, and demos of how to use PHILOW for various image types and segmentation targets.