http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-111192659-A

Outgoing Links

Predicate Object
assignee http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_6e11c1a8d75085dfc7964759844c563d
classificationCPCInventive http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G16H20-70
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G16H50-20
classificationIPCInventive http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G16H50-20
http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G16H20-70
filingDate 2019-12-31-04:00^^<http://www.w3.org/2001/XMLSchema#date>
inventor http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_d4a0d99e4c0568a7fd8cf010d2057266
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_4c316bc8a04053a0d01f1103372c39ce
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_128e2e471451eca441d4e022445385a4
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_ab1d5d0cc0ac718f16019240739f4f3e
publicationDate 2020-05-22-04:00^^<http://www.w3.org/2001/XMLSchema#date>
publicationNumber CN-111192659-A
titleOfInvention Pre-training method for depression detection and depression detection method and device
abstract The present invention discloses a pre-training method for depression detection and a depression detection method and device, wherein the method includes: dividing a spectrogram feature extracted from training audio into N sub-spectrogram features; Select k sub-spectrogram features before and after the center M 0 respectively, wherein, k<(N-1)/2; The first k sub-spectrogram features and the last k sub-spectrogram features of M 0 are collectively denoted as M i , and M i is input to the encoder; with the central sub-spectrogram feature M 0 as the target label, the encoder and decoder are trained to enable the encoder and the decoder to predict M 0 using M i . The solution provided by the method and device of the application can extract more abundant information about the human voice in the audio after pre-training the speech, so that the detection accuracy is greatly improved compared with no pre-training.
isCitedBy http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-114004973-A
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-113012720-A
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-112006697-A
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-113012720-B
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-112687390-A
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-112687390-B
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-113052113-B
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-113052113-A
priorityDate 2019-12-31-04:00^^<http://www.w3.org/2001/XMLSchema#date>
type http://data.epo.org/linked-data/def/patent/Publication

Incoming Links

Predicate Subject
isDiscussedBy http://rdf.ncbi.nlm.nih.gov/pubchem/substance/SID414881285
http://rdf.ncbi.nlm.nih.gov/pubchem/taxonomy/TAXID85571
http://rdf.ncbi.nlm.nih.gov/pubchem/anatomy/ANATOMYID85571
http://rdf.ncbi.nlm.nih.gov/pubchem/compound/CID17076

Total number of triples: 28.