Predicate |
Object |
assignee |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_ca46be4e68fa1f5a8b2ef4ab654c490a |
classificationCPCAdditional |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-10024 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-20084 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-20081 |
classificationCPCInventive |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06N3-08 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T5-003 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T5-005 |
classificationIPCInventive |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T5-00 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06N3-04 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06N3-08 |
filingDate |
2022-07-24-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
inventor |
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_9d0344aafcee54e1dd7a45fa0f49c23d http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_f6c8b4d6fb7cb0b7c8d9d94fae5beaa8 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_37f9356f0284cea85937b746b11b4de0 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_5de0b1cfcff427c1f652add8ee3c4f77 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_7945780de3d479516e998e86559e5d91 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_4ef8d101fc6042ab40c7563620291729 http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_82d47eb275a5e93ab735b2f3fbc00852 |
publicationDate |
2022-09-23-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
publicationNumber |
CN-115100076-A |
titleOfInvention |
A low-light image dehazing method based on context-aware attention |
abstract |
The invention discloses a low-light image dehazing method based on context-aware attention, which mainly solves the problems of poor generalization ability of the prior art model and serious color cast and unclear texture of the restored image. The scheme is: according to the MSBDNet framework, construct a low-light image dehazing network based on context-aware attention; take the acquired clear haze-free image set J t ' and clear low-light hazy image set It ' as the training image set; construct The joint total loss formula of the network; J t ', I t ' are equally divided into multiple paired image groups according to the batch size, and input 400 times in turn to the neural network to complete the training; input the image I c to be dehazed into The trained low-light dehazing neural network outputs a clear and haze-free image J c . The invention can restore the color tone and detail information of the image well, and its peak signal-to-noise ratio and structural similarity are both higher than or close to the prior art, and can be used for sharpening low-light foggy images. |
priorityDate |
2022-07-24-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
type |
http://data.epo.org/linked-data/def/patent/Publication |