Image Super Resolution
Image super-resolution is a machine learning task aimed at upscaling low-resolution images to high-resolution, typically with a magnification factor of 4x or higher, while preserving the content and details of the image as much as possible. This technology can significantly improve image quality, enhance visual effects, and increase the accuracy of computer vision algorithms, making it applicable to a wide range of scenarios.
ML-CrAIST
ABPN
WaveMixSR-V2
HAT
Config (e)
DRLN+
DBPN-RES-MR64-3
DBPN-RES-MR64-3
DBPN-RES-MR64-3
Edge-informed SR
IDM
DIP-HyperKite (ours)
ABPN
AESOP
RFB-ESRGAN
Ours w/o cycle-loss
RAMS (ours)
HiFaceGAN
HiFaceGAN
SROOE
FxSR-PD t=0.8
DAVI
EDSR+MMHCA
PASSRnet
PASSRnet
ABPN
DRCT-L
HMA†
SwinIR
DBPN-RES-MR64-3
PASSRnet
RankSRGAN
ATD
DRCT-L
DRLN+
DRCT-L
DBPN-RES-MR64-3
HAT-L
HMA†
DBPN-RES-MR64-3
StableShip
Extracter-rec
ABPN
DRLN+
HMA†
DRLN+
SRDRM-GAN
Full-GWAInet
GFRNet
ArcFace (0.4)