Electroretinogram Analysis Using a Short-Time Fourier Transform and Machine Learning Techniques

Faisal Albasu, Mikhail Kulyabin, Aleksei Zhdanov, Anton Dolganov, Mikhail Ronkin, Vasilii Borisov, Leonid Dorosinsky, Paul A. Constable, Mohammed A. Al-masni, Andreas Maier

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)
62 Downloads (Pure)

Abstract

Electroretinography (ERG) is a non-invasive method of assessing retinal function by recording the retina’s response to a brief flash of light. This study focused on optimizing the ERG waveform signal classification by utilizing Short-Time Fourier Transform (STFT) spectrogram preprocessing with a machine learning (ML) decision system. Several window functions of different sizes and window overlaps were compared to enhance feature extraction concerning specific ML algorithms. The obtained spectrograms were employed to train deep learning models alongside manual feature extraction for more classical ML models. Our findings demonstrated the superiority of utilizing the Visual Transformer architecture with a Hamming window function, showcasing its advantage in ERG signal classification. Also, as a result, we recommend the RF algorithm for scenarios necessitating manual feature extraction, particularly with the Boxcar (rectangular) or Bartlett window functions. By elucidating the optimal methodologies for feature extraction and classification, this study contributes to advancing the diagnostic capabilities of ERG analysis in clinical settings.
Original languageEnglish
Article number866
Number of pages21
JournalBioengineering
Volume11
Issue number9
DOIs
Publication statusPublished - Sept 2024

Keywords

  • electroretinography
  • biomedical signal processing algorithms
  • short-time Fourier transform
  • spectrogram
  • feature extraction
  • classification
  • machine learning
  • deep learning
  • neural network
  • retinal study

Fingerprint

Dive into the research topics of 'Electroretinogram Analysis Using a Short-Time Fourier Transform and Machine Learning Techniques'. Together they form a unique fingerprint.

Cite this