Project Description

Our approach for next-generation mobile health (mHealth) is centered in developing simple, yet reliable, technologies in which data-driven approaches minimize complicated hardware components and integrate into existing EHR systems. From a practical standpoint, it would not be possible to incorporate a bulky optical component in a conventional smartphone. Thus, computational spectrometry and imaging can play an important role in facilitating the development of mHealth technologies that can be embedded into conventional smartphones without any additional attachments. We are one of the first groups of working on ‘spectral super-resolution’ or ‘spectral reconstruction’ for medical and biological applications. For blood hemoglobin measurements, our group has developed hyperspectral reconstruction from RGB image data. Indeed, this is an interesting problem for machine learning and deep learning. Such a computational spectroscopy that does not require any accessory attachments to a conventional smartphone can potentially be scaled up to be integrated with EHR systems. We envision that mHeath technologies will allow us to leverage big clinical data sources (e.g. generated from EHR systems) and advanced data science tools (regression, compressive sensing, and machine learning) to improving health care and management for patients in resources limited settings, such as low- and middle-income countries and home settings.

Start Date 

2020

Postdoc Qualifications

Solid background and hands-on experience in statistics, compressive sensing, and machine learning. 

Solid background and hands-on experience in optical imaging and photonics. 

Co-advisors 

Young L. Kim
youngkim@purdue.edu
Weldon School of Engineering
http://web.ics.purdue.edu/~kim50/publication.htm

Paul M. Griffin
griff200@purdue.edu
Regenstrief Center for Healthcare Engineering
Industrial Engineering 
https://engineering.purdue.edu/IE/people/ptProfile?resource_id=159778 

Collaborator 

Martin Were, MD
Vanderbilt University School of Medicine and Moi University in Kenaya

References 

Virtual Hyperspectral Imaging of Eyelids - mHematology for Blood Hemoglobin Analysis
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3369797

Data-driven imaging of tissue inflammation using RGB-based hyperspectral reconstruction toward personal monitoring of dermatologic health
https://www.osapublishing.org/boe/abstract.cfm?uri=boe-8-11-5282