Active Tuberculosis Detection On CXR Package for Bob

Package to train convolutional neural networks for tuberculosis detection on frontal chest X-rays. Additionally, this package implements prediction of TB on chest X-rays and evaluation of performances. It is build using PyTorch.

User Guide

• Setup
  • Datasets
• Usage
  • Direct detection
  • Indirect detection
  • Scripts
• Results
  • Models optimization
  • Models training runtime and memory footprint
  • AUROC Scores
  • Global sensitivity analysis (relevance)
  • Ablation study
• References
• Supported Datasets
  • Tuberculosis datasets
  • Tuberculosis + radiological findings dataset
  • Radiological findings datasets
  • HIV-Tuberculosis datasets
• Command-Line Interface (CLI)
  • Setup
  • Preset Configuration Resources
  • Applications for experiments
• Application Program Interface (API)
  • Data Manipulation
  • Datasets
  • Engines
  • Neural Network Models
  • Toolbox
  • Preset Configurations

Indices and tables

• Index

• Module Index

• Search Page