Conţinutul numărului revistei |
Articolul precedent |
Articolul urmator |
1021 22 |
Ultima descărcare din IBN: 2024-03-15 14:26 |
SM ISO690:2012 CAPRITA, Horia, NEAMŢU, Mihai-Bogdan, MATACUTA, Ioana, BODRUG, Nicolae, DOBROTĂ, Luminiţa, NEAMŢU, Mihai-Leonida. Interpretarea semnalelor emg utilizând reţele
neuronale artificiale
. In: Revista ştiinţifico-practică ”Info-Med” , 2013, nr. 1(21), pp. 36-40. ISSN 1810-3936. |
EXPORT metadate: Google Scholar Crossref CERIF DataCite Dublin Core |
Revista ştiinţifico-practică ”Info-Med” | |||||
Numărul 1(21) / 2013 / ISSN 1810-3936 | |||||
|
|||||
Pag. 36-40 | |||||
|
|||||
Descarcă PDF | |||||
Rezumat | |||||
Human computer interaction is a rapid developing field. Human biosignals can be interpreted by computers (processors) working after mathematical algorithms developed by neurophysiologists, mathematicians and computer scientists. The perceptron(a single layer artificial network) has a supervised learning algorithm capability. Multilayer neural
network is a system of perceptrons with an Error Back Propagation learning algorithm which adjusts the sinaptic weights. A special type of neural networks is represented by
SOFM-Self organizing feature Map consisting of 2 layers of interconnections). For each neuron there is a cluster of vectors for inputs and one output. Human biosignal (EMG,EEG,EOG) have variables patterns and represents input vectors for these which learn to recognize a biosignal and transmit this in a specialised programme of a computer.
The article is a review regarding the accuracy of different types of neural networks (Back propagation neural network, Log-Linearised Gaussian Mixture Network, Fuzzy Mean Max Neural Network, Radial Basis Function Artificial Neural Network, Hidden Markov Model, Bayes networks) to classify the EMG signal and the capability to transmit the results to the main computer. |
|||||
Cuvinte-cheie human biosignals, EOG, EMG, neural networks., EEG |
|||||
|