Number of the records: 1
D3Net: Joint Demosaicking, Deblurring and Deringing
- 1.0537610 - ÚTIA 2022 RIV US eng C - Conference Paper (international conference)
Kerepecký, Tomáš - Šroubek, Filip
D3Net: Joint Demosaicking, Deblurring and Deringing.
2021 25th International Conference on Pattern Recognition (ICPR). Piscataway: IEEE, 2021, s. 8283-8290. ISBN 978-172818808-9. ISSN 1051-4651.
[2021 25th International Conference on Pattern Recognition (ICPR). Milan (IT), 10.01.2021-15.01.2021]
R&D Projects: GA ČR GA20-27939S
Grant - others:AV ČR(CZ) AP1701
Program: Akademická prémie - Praemium Academiae
Institutional support: RVO:67985556
Keywords : demosaicking * deblurring * deringing * ADMM * CNN
OECD category: Robotics and automatic control
http://library.utia.cas.cz/separaty/2021/ZOI/kerepecky-0537610.pdf
We have proposed a light-weight neural network architecture D3Net for joint demosaicking, deblurring and deringing by unrolling the ADMM optimization method.
Permanent Link: http://hdl.handle.net/11104/0316066
Number of the records: 1