A breakdown on some of the various projects I have done this last year.
Building.
Currently working towards solving the brain bandwidth problem and taking control of our cognitive evolution by
Learning.
Currently working towards solving the brain bandwidth problem and taking control of our cognitive evolution by
Building.
Currently working towards solving the brain bandwidth problem and taking control of our cognitive evolution by
The major difficulty in using EEG data is the variety of noises it produces. Current strategies forĀ findingĀ anomalies employ data annotation, which is subpar. In this research, the clustering technique allows us to classify artefacts in the preprocessing stage for cleaner data using K Means++.
A P300 spellerĀ is a communication tool with which one can input texts or commands to a computer by thought. The amplitude of theĀ P300Ā evoked potential is inversely proportional to the probability of infrequent or task-related stimulus. In tis project, I use the Farewell-Donchin paradigm to build the speller. The Farwell-Donchin paradigm has been a benchmark in P300 BCI. In this paradigm, a 6x6 matrix of letters and numbers is displayed and subject focuses on a target character while rows and columns of characters flash. By detecting P300 for one row and one column, the target character can be identified.
This study usesĀ convolutional neural network for classifying P300 electroencephalogram data. The Principal Component Analysis methodology was applied to both reduce noise and artefacts from the data and speed up the data processing. Using parallel convolution enhanced the classic convolutional neural network, enhancing the network's depth and capability to recognise P300 electroencephalogram signals. Accuracy rates obtained were greater than 88%.
In this project, we propose brain "bookmarking" based on situational interest. We designed systems for digital saving based on interest and what you want to reference back to for learning/deeper exploration to enable faster pace at workplace. In order to evaluate the situational interest of an individual, the EEG data was preprocessed and then decomposed using Empirical Mode Decomposition (EMD). For each participant, a matrix was created utilising the best six characteristics from four EEG channels. Support Vector Machine (SVM) classifiers were fed these chosen characteristics.
Every minute, 6 people die due to an alcohol disorder. In this project, I used a KNN classifier to correlate specific EEG signals with alcoholic subjects with a 96% accuracy. Beta and Gamma waves are used to observe and classify the data.
One of many biometric technologies is being offered as a viable alternative to passwords. The concept proposes that a person's identity may be validated by EEG data. We observed that the EEG "signatures" are unique and more sophisticated than a normal password, making them harder to hack.
We developed our own exclusive 3-in-1 technology that allows us to diagnose, stimulate, and cure any brain problems connected with the limbic system. We take three stages: nanoparticle binding, neurostimulation (Magnetogenetics), and magnetic particle imaging.
MikaelĀ Haji Ā©Ā 2021. All rights reserved.