Biomedical Signal Classification Using Fuzzy Classifiers

Project Description

Project number: 09I03-03-V04-00416

Web page: https://ki.fri.uniza.sk/biosig/index.html

Signal classification is often used in many industrial applications and functioning is based on the result of automatic classification of signals. It causes strict requirements for classification accuracy and the development of new methods and algorithms of highly accurate signal classification. One of the areas with higher requirements for signal classification accuracy is medicine. Biomedical signal processing involves acquiring and preprocessing physiological signals and extracting meaningful information to identify patterns and trends within the signals. Sources of biomedical signals include neural activity, cardiac rhythm, muscle movement, and other physiological activities. For example, typical biomedical signals focused on neural activity are based on Electroencephalogram (EEG) analysis. This signal is commonly used in the development of human-computer interaction and medicine. Medical applications of automatic EEG classification are known for the diagnosis of epilepsy, Alzheimer's disease, depression, and other diseases. Every one of these applications in medical diagnosis has some specifics related to different properties of EEG signals. The most investigated EEG diagnosis application is the diagnosis of epilepsy, which belongs to the most common chronic neurological disorders.

A Major Objectives

• Bio-signal Pre-Processing: Development of procedures for preprocessing. Developed methods will address how to deal with uncertain data. One possible solution is the application of fuzzy logic.

• Bio-signal Classification: Development of whole classification approach for medical signal classification. The efficient classification of uncertain data often requires the application of fuzzy classifiers. This type of classifier allows for the increasing accuracy of classification for uncertain data.

• Open-source Software: Development of open-source software that will integrate developed methods and approaches. The application will work with a user interface in which the user can make the design of the data mining process. The application will also be used in two courses at Faculty of Management Science and Informatics.

  • 

• Result Integration: The project results will be made available through open science principles. The developed software will incorporate the methods created and will be used in two courses at the Faculty of Management Science and Informatics.

Project Outputs: