Development and validation of a paper-based sensor and convolutional neural networks integrated in a single-board computer for the quantification of sodium iron-EDTA and zinc oxide in fortified corn flours

Micronutrient deficiencies have a high prevalence in low- and middle-income countries (LMICs), particularly among children under five years and women of reproductive age. To combat micronutrient deficiencies in LMICs, the addition of micronutrients to fortify staple foods, e.g., corn flour, through rural small-scale mills has proven to be a cost-effective strategy. The laboratory capacity to assess quality control and assurance (QC/QA) in these contexts, however, is either unavailable or prohibitively expensive. The overall objective of this dissertation is to design, develop, and validate two paper-based sensors which interface with a built-in-house computer vision-based analyzer to quantify iron and zinc in corn flour. The dissertation will accomplish the aforementioned objective using mixed-methods studies. The first study will develop a paper-based sensor to quantify iron concentration in fortified corn flour and will validate its performance using samples from Tanzania’s small-scale mill fortification programs to determine its applicability as a monitoring and evaluation tool. In addition, this study will optimize the mineral extraction method and sampling instrument design. The second study will develop a paper-based sensor to quantify zinc concentration in fortified corn flour and will validate its performance using samples from Tanzania’s small-scale mill fortification programs to determine its applicability as a means of monitoring and evaluating. The third study will design and optimize a convolutional neural network architecture as the backbone of the computer vision-based analyzer that predicts iron and zinc concentration using digital images from the paper-based sensor. In collaboration with SANKU Project Healthy Children, a non-governmental organization supporting flour fortification in Tanzania, this dissertation ultimately serves as a bottom-up solution to the monitoring and evaluation limitations of low-resource fortification programs.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Food Science & Human Nutrition

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Toc Sagra, Marco Eduardo

Contributors dc:contributor

• Andrade, Juan E.
• Stasiewicz, Matthew J.
• Engeseth, Nicki J.
• Wang, Yi-Cheng

• Paper-based Sensor
• Computer Vision
• Quality Control

Statement dc:rights

• Copyright 2023 Marco Eduardo Toc Sagra

Language dc:language

eng

Handle dc:identifier

https://hdl.handle.net/2142/121201

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F121201