Development of a Web Application for Audio Communication using EEG Signals

An accurate analysis of electroencephalogram (EEG) data was achieved to identify the needs of the patients who are unable to coherently articulate their needs to caretakers. An audio output via Bluetooth speaker was designed to articulate on behalf of the patient to communicate his/her needs to the caretaker. The analysis done in this project was on the cognitive stimuli-based EEG data. The project goes on in three main stages, Firstly, making of the web application with user friendly interface; Secondly, building of the machine learning model to identify the need of the patients and thirdly, provide an audio output suited for the identified need, so that the caretaker will know what exactly the patient need. The User interface is made using html and CSS and bootstrap is used to make the application more responsive and dynamic. The main pages contain the links that can be used to access the different parts of the project. Each page is supported by a python backend and it is applied using the flask python framework. The machine learning models are pre-trained and stored using pickle in python. The models were trained with the cognitive dataset that was provided and the performances of the models are recorded. A training section is also provided in the web application that can be used to demonstrate the performance of the model which will display the accuracy score of the trained data. Each time a prediction is made in the algorithm page the python backend is programmed so that the output of the prediction is given back to the caretaker in audio form