Predicate |
Object |
assignee |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_b741ec085646f4ce9c814a01617717e8 |
classificationCPCAdditional |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-10048 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-20084 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-20081 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-10016 |
classificationCPCInventive |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-174 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-215 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-194 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-38 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-337 |
classificationIPCInventive |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-38 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-174 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-33 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-215 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-194 |
filingDate |
2020-11-04-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
inventor |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_3ef3a6b77aa1b2b35f65222f7905dea8 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_61a72693a5ac456f112ea4b0614bdd58 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_84510a3a544f12ef56173f333a70b7b3 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_561fa07bd292c12deff417d8a040ecc3 |
publicationDate |
2021-01-26-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
publicationNumber |
CN-112270688-A |
titleOfInvention |
Foreground extraction method, device, equipment and storage medium |
abstract |
The application relates to a foreground extraction method, a device, equipment or a storage medium, wherein the method comprises the following steps: acquiring an infrared image sequence and a visible light image sequence; the infrared image sequence corresponds to the visible light image sequence one by one; carrying out contrast processing on each frame of infrared image in the infrared image sequence to obtain an initial mask image sequence; connecting the initial mask image sequence and the visible light image sequence in series to obtain an input image sequence; performing feature extraction on an input image sequence according to the trained feature extraction model to obtain a feature tensor; determining a prediction mask image corresponding to the characteristic tensor according to the trained long-time memory network model; and determining a foreground image according to the predicted mask image and the visible light image. The foreground extraction method provided by the embodiment of the application is suitable for extracting static or moving foreground objects, including semitransparent objects, motion fuzzy objects and virtual focus objects, and can improve the extraction precision of the edges of the foreground objects. |
isCitedBy |
http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-113822818-A http://rdf.ncbi.nlm.nih.gov/pubchem/patent/TW-I792978-B http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-113762253-A |
priorityDate |
2020-11-04-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
type |
http://data.epo.org/linked-data/def/patent/Publication |