BDIS: Balanced Training Architecture for Dual Image Scaler Using Origin Referenceable Losses

BDIS: Balanced Training Architecture for Dual Image Scaler Using Origin Referenceable Losses

Deep neural network (DNN)-based research on image scaling has mostly focused on super-resolution (SR) rather than image downscaling. Specifically, most existing DNN-based methods for image downscaling are used as auxiliary modules to improve the quality of super-resolved images. In rare cases, DNN-b...

Full description

Bibliographic Details
Main Authors: Eun Su Kang, Jung Eun Kwon, Hae Ju Park, Moon Ju Chae, Sung In Cho
Format: Article
Language:English
Published: IEEE 2022-01-01
Series:IEEE Access
Subjects:
Balanced architecture
deep neural network
image downscaling
image upscaling
image super-resolution
dual scaler
Online Access:https://ieeexplore.ieee.org/document/9780200/

Internet

https://ieeexplore.ieee.org/document/9780200/

Similar Items