Accès libre

Research on the Reconstruction Technology of Digitized Artworks Based on Image Processing Algorithms and Its Cultural Inheritance Value

  
25 sept. 2025
À propos de cet article

Citez
Télécharger la couverture

Figure 1.

Change of the consistent loss of the cycle
Change of the consistent loss of the cycle

Figure 2.

Change of significant edge loss
Change of significant edge loss

Figure 3.

The clarity of the generated image
The clarity of the generated image

Image quality assessment generated by different algorithms

Algorithm SSIM COSIN
CycleGAN 0.212 0.834
Algorithm of this article 0.442 0.97
DaulGAN 0.224 0.937
Pix2pix 0.422 0.96

Eye tracking data

Method Categories of artworks Name Fixation time(s) Number of watch
Method of this article Painting works of art The last dinner 14.85 62
Night tour 9.49 44
The Virgin of Rocks 14.54 75
Sculpture art David 18.83 92
Thinkers 16.26 64
Venus ‘broken arm 14.22 67
Craft artwork Terracotta Warriors 14.31 67
Tang tricolor 12.73 71
Blue and white porcelain 18.77 32
Average 14.89 64
Pix2pix Painting works of art The last dinner 11.04 20
Night tour 13.23 23
The Virgin of Rocks 13.3 37
Sculpture art David 8.69 5
Thinkers 6.76 8
Venus ‘broken arm 10.15 34
Craft artwork Terracotta Warriors 9.44 29
Tang tricolor 8.24 5
Blue and white porcelain 7.17 18
Average 9.78 20

Artworks

Categories of artworks Name
Painting works of art The last dinner
Night tour
The Virgin of Rocks
Sculpture art David
Thinkers
Venus ‘broken arm
Craft artwork Terracotta Warriors
Tang tricolor
Blue and white porcelain

The classification accuracy of images

Data set Monet2photo Horse2zebra Summer2winter
Real image 0.93 0.91 0.88
μ =0 0.85 0.81 0.8
μ =1 0.85 0.74 0.82
μ =5 0.9 0.87 0.9
μ =10 0.89 0.79 0.87
μ =20 0.91 0.72 0.73
μ =50 0.78 0.85 0.8