A novel pilot study of automatic identification of EMF radiation effect on brain using computer vision and machine learning

Authors: Singh A, Singh N, Jinda T, Rosado-Muñoz, A, Kishore Dutta M

Year: March 2020

Category: Biomedical Signal Processing and Control

Journal: Biomedical Signal Processing and Control

DOI: https://doi.org/10.1016/j.bspc.2019.101821

URL: https://www.sciencedirect.com/science/article/abs/pii/S1746809419304021

Abstract

Overview

This study explores the impact of Electromagnetic Field (EMF) radiation, emanating from sources such as mobile phones and cell towers, on the brain. It highlights the neurological changes that can occur due to EMF radiation, evidencing both morphological and chemical alterations in the brain and other internal organs.

Methodology

  • The study utilized Drosophila melanogaster as a model organism.
  • It involved automatic brain segmentation and microscopic imaging to analyze brain morphology changes under EMF exposure.
  • Discriminatory geometrical features from these images were then extracted and analyzed.

Findings

Significant discriminatory features in the brain's morphology were identified and used successfully in machine learning models to distinguish between EMF-exposed and non-exposed specimens. Four different classifiers were used:

  • Support Vector Machine
  • Naïve Bayes
  • Artificial Neural Network
  • Random Forest

Classification accuracy of up to 94.66% was achieved using features selected through a feature selection method, suggesting that the approach is efficient and yields reliable diagnostic results with low time complexity.

Conclusion

This pilot study offers a novel, automated, and time-efficient method to detect EMF exposure effects on the brain using advanced image processing and machine learning techniques.

EMF brain machine learning computer vision neuroscience EMF exposure image processing
← Back to Stats