Deep Phenotyping Documentation

Contrastive Representation Learning for Single Cell Phenotyping in Fluorescent Whole Slide Imaging

This documentation provides comprehensive information about the Deep Phenotyping framework, which uses contrastive learning for identifying and stratifying single cells in whole slide immunofluorescence microscopy images derived from liquid biopsies. The framework achieves high accuracy in classifying diverse cell phenotypes and enables automated identification of rare cell populations.

Core Modules

Representation Learning

Core implementation of the contrastive learning framework for feature extraction from cell images.

Contrastive Learning Feature Extraction Neural Networks

Leukocyte Classifier

Classification models for leukocytes and other cell types in liquid biopsies.

Classification Cell Types Machine Learning

Utilities

Common utilities for data processing, normalization, and image analysis.

Image Processing Data Handling Helper Functions

Pipeline Components

Pipeline

End-to-end processing pipeline for whole slide images, from preprocessing to analysis.

Workflow Processing Analysis