HyperAI

Image Super Resolution On Bsd100 4X Upscaling

Metrics

PSNR
SSIM

Results

Performance results of various models on this benchmark

Comparison Table
Model NamePSNRSSIM
activating-more-pixels-in-image-super28.050.7534
lightweight-and-efficient-image-super27.570.7353
a-fully-progressive-approach-to-single-image27.79-
drct-saving-image-super-resolution-away-from28.160.7577
structure-preserving-super-resolution-with25.5050.6576
zero-shot-super-resolution-using-deep27.120.7211
photo-realistic-single-image-super-resolution25.160.6688
image-super-resolution-with-cross-scale-non27.80.7439
beyond-deep-residual-learning-for-image27.660.7380
photo-realistic-single-image-super-resolution27.580.762
drct-saving-image-super-resolution-away-from28.060.7533
non-local-recurrent-network-for-image27.480.7306
lightweight-image-super-resolution-with-127.56-
one-to-many-approach-for-improving-super--
residual-dense-network-for-image-super27.720.7419
channel-partitioned-windowed-attention-and28.060.7532
progressive-perception-oriented-network-for27.83-
gated-multiple-feedback-network-for-image27.740.7421
wavemixsr-v2-enhancing-super-resolution-with27.870.764
image-super-resolution-via-rl-csc-when27.440.7302
auto-encoded-supervision-for-perceptual-image25.930.6813
single-image-super-resolution-via-a-holistic27.850.7454
enhancenet-single-image-super-resolution27.500.7326
photo-realistic-single-image-super-resolution25.940.6935
a-framework-for-real-time-object-detection27.760.7441
multi-level-wavelet-cnn-for-image-restoration27.620.7355
feedback-network-for-image-super-resolution27.720.7409
flexible-style-image-super-resolution-using24.770.6817
local-texture-estimator-for-implicit27.86-
channel-partitioned-windowed-attention-and28.040.7527
real-time-single-image-and-video-super27.020.7442
densely-residual-laplacian-super-resolution27.870.7453
flexible-style-image-super-resolution-using26.380.738
second-order-attention-network-for-single27.860.7457
joint-maximum-purity-forest-with-application26.87-
recovering-realistic-texture-in-image-super25.330.651
sesr-single-image-super-resolution-with27.420.737
image-super-resolution-using-very-deep27.770.7436
image-restoration-using-very-deep27.400.7290
image-restoration-using-deep-regulated27.21-
enhanced-deep-residual-networks-for-single27.710.7420
hierarchical-back-projection-network-for27.770.743
image-super-resolution-using-deep26.90.7101
image-super-resolution-via-attention-based27.820.743
learning-a-single-convolutional-super27.490.734
edge-informed-single-image-super-resolution24.250.851
photo-realistic-single-image-super-resolution25.020.6606
beyond-a-gaussian-denoiser-residual-learning27.290.7253
ram-residual-attention-module-for-single27.560.7350
hierarchical-information-flow-for-generalized28.130.7622
image-super-resolution-via-dual-state27.250.724
deep-laplacian-pyramid-networks-for-fast-and27.320.728
seven-ways-to-improve-example-based-single27.16-
deeply-recursive-convolutional-network-for27.210.7493
image-restoration-using-convolutional-auto27.40.729
hmanet-hybrid-multi-axis-aggregation-network28.130.7562
perception-oriented-single-image-super24.870.6869
memnet-a-persistent-memory-network-for-image27.400.7281
swinfir-revisiting-the-swinir-with-fast28.07-
lightweight-feature-fusion-network-for-single27.42-
esrgan-enhanced-super-resolution-generative27.850.7455
wavemixsr-a-resource-efficient-neural-network-0.7605
progressive-perception-oriented-network-for-0.7515
deep-back-projection-networks-for-super27.720.740
image-super-resolution-via-feature-augmented26.91-
activating-more-pixels-in-image-super28.090.7551
deep-learning-based-image-super-resolution26.57070.6900
fast-accurate-and-lightweight-super-127.580.7349
fast-and-accurate-single-image-super27.410.7297
swinfir-revisiting-the-swinir-with-fast28.030.7520
perceptual-losses-for-real-time-style24.950.6317