Applied Chest Imaging Laboratory


The Applied Chest Imaging Laboratory is committed to the dissemination of open and reproducible science. By fostering collaboration and knowledge-sharing, we aim to accelerate scientific discoveries and improve patient outcomes in the realm of chest imaging and pulmonary health. 

The Applied Chest Imaging Laboratory continuously updates and refines its software packages to incorporate the latest advancements in imaging technology and algorithms, as well as to address the evolving needs of our user community. We welcome feedback and contributions from users and collaborators, as we believe that collective expertise and a shared commitment to open science will drive innovation and excellence in the field.

Chest Imaging Platform:

Open-source platform to phenotype chest CT images. CIP is part of the NHLBI BioData Catalyst tools ecosystem, designed to enable large-scale phenotyping. CIP also offers a user-friendly workstation environment for research purposes, available as a 3D Slicer Extension, facilitating seamless integration into your workflow.

Website   Github   Docker 

Bayesian Trajectory Modeling:

Advanced Bayesian Trajectory Modeling Python Package: Our innovative software solution harnesses cutting-edge advancements in Bayesian modeling to deliver unparalleled trajectory modeling capabilities. With an easy-to-use Python package, researchers can effortlessly integrate our unique methods into their discovery pipelines, elevating the quality and precision of their trajectory analyses.

Github  PiP

CXR and COVID-19:

Mobile Solution for Mild Pneumonia Detection in Various Settings.¬†The Applied Chest Imaging Lab introduced the ‘SlowDown COVID-19’ initiative in Spring 2020, delivering rapid diagnostic solutions at a critical time. This mobile tool is designed to quickly and accurately detect mild pneumonias, supporting healthcare providers in managing the pandemic across diverse environments.

Website  GitHub  MobileApp

Distributed Hyperparam Tuning:

Distributed Hyperparameter Tuning with Spearmint.¬†Enhance your algorithm’s parameter tuning process using Bayesian optimization. Our open-source client/server implementation, based on Spearmint, is designed to streamline and optimize your parameter tuning tasks, leveraging the power of Bayesian optimization for improved performance and efficiency.

Server Github  Client Github  Web Service

Thorax Segmentation in CXR:

Thoracic Segmentation in CXR. Efficiently isolate thoracic and mediastinal regions in chest X-rays for focused analysis using our open-source segmentation model. This reliable solution has been validated against expert opinion, ensuring accurate delineation of key anatomical structures to enhance your diagnostic workflow.