Presenter: Delan Sherzad Mohammed Salih
Host:
Title of the Seminar: Electrogastrogram (EGG) data augmentation

Objectives or Summary:
1. Address Dataset Limitations:
To tackle the scarcity and imbalance of publicly available Electrogastrography
(EGG) datasets, which hinder the development of robust analytical frameworks,
particularly for Machine Learning (ML) applications.
2. Enhance Dataset Size and Balance:
To employ signal processing techniques and data augmentation methods (e.g.,
mirroring, brightness adjustment, and circular shifting) to expand and balance a
limited EGG dataset, ensuring a more representative and unbiased dataset for ML
model training.
3. Preserve Signal Integrity:
To ensure that the essential frequency characteristics of the original EGG signals are
preserved during augmentation, maintaining the diagnostic relevance of the data.
4. Enable ML-Compatible Data Representation:
To convert 1-dimensional EGG signals into 2-dimensional spectrograms using
Continuous Wavelet Transform (CWT), facilitating image-based ML inputs and
improving the potential for accurate signal classification.

Place of the Seminar: ECE seminar hall