Related Experiment Video
Updated: Jun 12, 2025
13:19
Deep Neural Networks for Image-Based Dietary Assessment
Published on: March 13, 2021
9.0K
Smartphone-Assisted Nanozyme Colorimetric Sensor Array Combined "Image Segmentation-Feature Extraction" Deep Learning
Xinyu Zhong1, Yuelian Qin1, Caihong Liang2
1Pharmaceutical College, Guangxi Medical University, Nanning 530021, China.
ACS Sensors
|September 19, 2024
Summary
This study introduces a low-cost smartphone sensor array and deep learning for rapid detection of unsaturated fatty acids (UFAs). The intelligent platform accurately identifies and quantifies UFAs in oils, overcoming limitations of traditional methods.
Area of Science:
- Analytical Chemistry
- Materials Science
- Biotechnology
Background:
- Conventional methods for unsaturated fatty acid (UFA) detection are often time-consuming, require extensive sample preparation, and rely on costly instrumentation.
- There is a need for rapid, cost-effective, and portable analytical tools for UFA analysis in various applications, including food quality control.
Purpose of the Study:
- To develop an intelligent platform for facile and low-cost qualitative and quantitative analysis of unsaturated fatty acids (UFAs).
- To combine a smartphone-assisted colorimetric sensor array (CSA) with deep learning for enhanced UFA detection capabilities.
Main Methods:
- Development of a CSA utilizing MnO2 nanozymes, enhanced by doping with Ag, Pd, and Pt, mimicking the mammalian olfactory system.
- Implementation of an "image segmentation-feature extraction" deep learning (ISFE-DL) approach using the MobileNetV3 small model for multicomponent quantitative analysis.
- Utilized smartphone apps "Quick Viewer" and "Intelligent Analysis Master" for data acquisition, processing, and one-click quantification.
Main Results:
- The CSA successfully discriminated between oleic acid (OA), linoleic acid (LA), and α-linolenic acid (ALA), including their mixtures and various edible oils.
- High determination coefficients were achieved for OA (0.9969), LA (0.9668), and ALA (0.7393) using the ISFE-DL model.
- The platform demonstrated effective differentiation of camellia oils (CAO) and detection of adulterated samples.
Conclusions:
- The developed intelligent platform offers an innovative and efficient approach for the rapid qualitative and quantitative analysis of UFAs.
- The combination of smartphone-based CSA and ISFE-DL provides a low-cost, portable, and sensitive alternative to conventional analytical methods.
- This technology holds potential for analyzing other compounds with similar characteristics, expanding its applicability.
