Resource type
Thesis type
(Thesis) M.Sc.
Date created
2025-08-06
Authors/Contributors
Author: Mirzaei, Shahrzad
Abstract
Most current methods for liquid analysis and fraud detection rely on expensive tools and controlled lab environments, making them inaccessible to most lay users. We present MobiLyzer, a mobile system that enables fine-grained liquid analysis using unmodified commodity smartphones in real-world settings. MobiLyzer performs spectral analysis by leveraging the light reflectance properties of liquids across wavelengths. Spectral analysis of liquids on smartphones is challenging due to limited sensing, camera variations, and uncontrolled ambient illumination. Diverse illumination introduces distortions in measured spectra; liquid containers cause specular reflections that degrade accuracy. MobiLyzer uses RGB images captured by smartphone cameras and intrinsic decomposition to standardize data across different illuminations and separates the illumination effect from the true colors of the liquid. It further leverages NIR sensors on phones to collect complementary images. A machine-learning model processes RGB and NIR images to reconstruct VNIR spectra, preserving spectral and color characteristics, critical as many liquids differ only in subtle spectral cues, enabling high-fidelity analysis without expensive cameras. MobiLyzer can detect low-concentration adulteration in visually similar substances, differentiate quality grades (e.g., refined vs. extra virgin olive oil), identify geographic origin (e.g., Italian vs. U.S. olive oil), and quantify concentrations (e.g., urine protein for early kidney disease detection).
Document
Extent
49 pages.
Identifier
etd23933
Copyright statement
Copyright is held by the author(s).
Academic Supervisor
Thesis advisor: Hefeeda, Mohamed
Language
English
Member of collection
| Download file | Size |
|---|---|
| etd23933.pdf | 32.85 MB |