http://rdf.ncbi.nlm.nih.gov/pubchem/patent/KR-20080009458-A
Outgoing Links
Predicate | Object |
---|---|
assignee | http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_142d981bd6c180968f769e0da8f620da |
classificationCPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/A61B5-165 |
classificationIPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/A61B5-04 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/A61B5-0476 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/A61B5-0484 |
filingDate | 2006-07-24-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
inventor | http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_44dacb6988097feffed4c00077f5b2a0 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_a5bef532ad592fa4af793713a0330332 |
publicationDate | 2008-01-29-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
publicationNumber | KR-20080009458-A |
titleOfInvention | Emotion Recognition Device Using Neural Network |
abstract | An emotion recognition system using a neural network is disclosed. The data generation unit calculates a ratio of the power value of each frequency signal to the sum of the power values of the signals belonging to the respective frequency bands divided by the predetermined frequency bands from the EEG signal measured from the subject, The power value corresponding to the frequency signal is quantified, and each of the quantified power values is changed to a preset level value according to the magnitude, thereby generating input data. The neural network is composed of an input layer into which input data is input, an output layer from which emotion values corresponding to the input data are output, and a hidden layer positioned between the input layer and the output layer, and between the input layer and the hidden layer, and between the hidden and output layers, It is connected to the connection strength, which is a correlation of emotion values, and compares emotion values output through the output layer with preset emotion values in response to input data, and changes the connection strength between layers based on the error rate between each layer. The learning unit inputs the input data corresponding to the EEG signal measured for each emotional state of each subject as learning data and inputs it to the input layer of the neural network, and connects when the error rate of the output layer reaches a preset reference error rate. The strength is determined by the connection strength of the neural network corresponding to each emotional state. According to the present invention, a more accurate recognition result can be obtained by determining the connection strength of neural networks by performing learning with quantified EEG signals according to the user's emotion. |
isCitedBy | http://rdf.ncbi.nlm.nih.gov/pubchem/patent/KR-101508200-B1 http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-110507335-A http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-110507335-B http://rdf.ncbi.nlm.nih.gov/pubchem/patent/KR-101325189-B1 http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-110141258-A http://rdf.ncbi.nlm.nih.gov/pubchem/patent/KR-20200083027-A |
priorityDate | 2006-07-24-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
type | http://data.epo.org/linked-data/def/patent/Publication |
Incoming Links
Total number of triples: 24.