Date of Award
Fall 11-21-2025
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Computer Science
First Advisor
Sheikh Ahamed
Second Advisor
Danny Thomas
Third Advisor
Praveen Madiraju
Abstract
Infectious diseases remain one of the leading global causes of illness and death in both community and hospital settings. Infections lead to lasting health issues, such as life-threatening chronic conditions or even cancer. Early detection is crucial for timely intervention, such as isolation and treatment. Hematology originated from the Greek word haima, meaning blood. It is a unique discipline in medicine that involves the study of blood and its contents. White blood cells (WBC) are at the forefront of the body’s immune system. WBCs are categorized into five main subtypes: neutrophils, lymphocytes, monocytes, eosinophils, and basophils. Monitoring the total WBC and its subtypes is critical for assessing immune function and diagnosing various health conditions. Granulocytes—comprising neutrophils, eosinophils, and basophils—play key roles in inflammation and infection response. In case of infections, WBC production is increased to fight off foreign pathogens. Although not part of the immune system, red blood cells (RBCs) can indirectly reflect immune activity, as inflammatory responses may alter iron metabolism and reduce hemoglobin levels. Monitoring hemoglobin changes can help assess disease severity, immune response, and recovery, making it a valuable biomarker in infection-related diagnostics and patient management. Conventional techniques, such as the Complete Blood Cell (CBC) count test requires blood samples and specialized equipment, limiting frequent assessments. Another prominent defense mechanism against infections is the body's ability to elevate its temperature to create an environment that hinders pathogen growth. This rise in temperature results in fever. The most common way of measuring core body temperature is using a thermometer, but using shared thermometers may pose a risk of cross-contamination in clinical settings. To overcome these limitations, this study introduces four innovative, smartphone-based, non-invasive solutions for monitoring hematologic and thermal biomarkers: 1. UbiWhite: A non-invasive OptoMagnetic system for real-time WBC counting from fingertip videos. 2. UbiWhite-Plus: A framework using optical properties and deep learning for non-invasive granulocyte detection from fingertip videos. 3. UbiHemo-AI: Quantifies hemoglobin levels accurately from a 10-second fingertip video captured under near-infrared lighting, augmented by generative AI. 4. UbiFaceTemp: A fully contact-free system estimating core body temperature from a 10-second face video using image processing and thermodynamics.
Comments
Doctor of Philosophy (PhD)