Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper explores the use of Convolutional Neural Networks (CNNs) and Adversarial Autoencoders (AAEs) for classifying EEG data related to hand movement. The approach involves converting raw EEG signals into 2D topograms and then training both supervised CNNs and semi-supervised AAEs to classify different motor cortex activities. The study demonstrates the feasibility of applying computer vision techniques to EEG data analysis for brain activity classification.
Computer vision techniques, typically used for images, can be successfully adapted to classify brain activity from EEG data.
In this paper, we consider applying computer vision algorithms for the classification problem one faces in neuroscience during EEG data analysis. Our approach is to apply a combination of computer vision and neural network methods to solve human brain activity classification problems during hand movement. We pre-processed raw EEG signals and generated 2D EEG topograms. Later, we developed supervised and semi-supervised neural networks to classify different motor cortex activities.