Image-Based Neural Network for the Classification of Paraquat Concentrations in Urine Samples

A computer vision and machine learning-based system was developed for the classification of paraquat concentrations in images of urine samples. This system employs a multilayer perceptron (MLP) fed by a feature vector extracted from each image based on the L*a*b* and HSV (Hue, Saturation, Value) color spaces. The vector values were determined through the median and standard deviation of three regions of interest (ROIs) located at different positions along the test tube. Through an ablation study, in which the depth and neuron density of the multilayer perceptron were varied and the results of the classification performance tests were compared, it was determined that the optimal neural network architecture consists of two hidden layers with 16 and 32 neurons, respectively. With this configuration, an accuracy of 98.06% was reached on an independent test subset.

Authors

Institutions

Publication Details

Journal
Technologies
Published
2026-11-17
DOI
https://doi.org/10.3390/technologies14110760
Primary Topic
Paraquat toxicity studies and treatments
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Image-Based Neural Network for the Classification of Paraquat Concentrations in Urine Samples

Maricela Jiménez Rodríguez, Joel Hinojosa-Dávalos, Cuauhtémoc Acosta Lúa, Joel Salazar‐Flores et al.
Technologies
Paraquat toxicity studies and treatments
article

Image-Based Neural Network for the Classification of Paraquat Concentrations in Urine Samples

Maricela Jiménez Rodríguez, Joel Hinojosa-Dávalos, Cuauhtémoc Acosta Lúa, Joel Salazar‐Flores, Carlos Alfonso Flores-Gutierrez, Ariadna Berenice Flores Jiménez
article en

Abstract

A computer vision and machine learning-based system was developed for the classification of paraquat concentrations in images of urine samples. This system employs a multilayer perceptron (MLP) fed by a feature vector extracted from each image based on the L*a*b* and HSV (Hue, Saturation, Value) color spaces. The vector values were determined through the median and standard deviation of three regions of interest (ROIs) located at different positions along the test tube. Through an ablation study, in which the depth and neuron density of the multilayer perceptron were varied and the results of the classification performance tests were compared, it was determined that the optimal neural network architecture consists of two hidden layers with 16 and 32 neurons, respectively. With this configuration, an accuracy of 98.06% was reached on an independent test subset.

TechnologiesVol. 14(11)
Universidad de Guadalajara (MX), Universidad de la Ciénega (MX)
Paraquat toxicity studies and treatments
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon