Table 2 Mean Opinion Scores (MOS) of different methods on the OBITrue dataset

From: Unsupervised GAN-based model for super-resolution of oracle bone rubbing images

Method

Image Clarity↑

Structural Consistency↑

Semantic Interpretability↑

Average MOS↑

Bicubic

1.5

2.1

3.0

2.20

CycleGAN

4.2

3.2

2.9

2.43

LRGAN

3.8

3.0

3.8

3.53

HiFaceGAN

2.4

2.2

3.2

2.60

MM-RealSR

2.1

1.8

2.9

2.27

SCGAN

4.3

1.6

2.7

2.87

WFEN

4.1

4.3

4.5

4.30

NeXtSRGAN

2.5

1.7

2.8

2.33

OBISR-DTG

4.6

4.4

4.6

4.53

OBISR-DTGEMA

4.4

4.2

4.5

4.36

  1. Scores are averaged over 16 participants, comprising both domain experts and graduate students, based on image clarity, structural consistency, and semantic interpretability. In the table, OBISR-DTG and OBISR-DTGEMA denote the outputs generated by the DTG and DTGEMA generators within the OBISR model.
Back to article page