Small Sample Bronze Inscription Classification Algorithm Based on Morphological Features

Tongyao Luo; Huiqin Wang; Ke Wang; Zhan Wang; Hong Wang

doi:10.3788/LOP212983

Laser & Optoelectronics Progress, Volume. 60, Issue 4, 0410011(2023)

Small Sample Bronze Inscription Classification Algorithm Based on Morphological Features

Tongyao Luo¹, Huiqin Wang^1、*, Ke Wang¹, Zhan Wang², and Hong Wang²

¹School of Information and Control Engineering, Xi'an University of Architecture and Technology, Xi'an, Shaanxi 710055, China

²Shaanxi Provincial Institute of Cultural Relics Protection, Xi'an 710075, Shaanxi, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(11)

Fig. 1. Classic AlexNet model

Download full size

Fig. 2. Determination of main direction

Download full size

Fig. 3. Rubbings of primitive bronze inscriptions

Download full size

Fig. 4. Rubbings of repaired bronze inscriptions

Download full size

Fig. 5. Structure diagram of AlexNet model integrating morphological details of SURF inscription

Download full size

Fig. 6. Bronze inscription picture pre-processed. (a) Picture of bronze inscriptions in the New Inscriptions on Bronze Chronicles;(b) pictures of inscription after cropping; (c) bronze inscription pictures after hole filling

Download full size

Fig. 7. Comparison of inscription classification results before and after algorithm improvement. (a) AlexNet_loss; (b) AlexNet_accuracy; (c) Improved_AlexNet_loss; (d) Improved_AlexNet_accuracy; (e) proposed_loss; (f) proposed_accuracy

Download full size

Fig. 8. Real inscription picture

Download full size

Fig. 9. Classification and recognition results of bronze inscriptions

Download full size

Table 1. Classification accuracy of different algorithms
View table
Table 1. Classification accuracy of different algorithms
Five-fold cross validation Algorithm accuracy /%
AlexNet Improved_AlexNet SURF Improved_AlexNet+SERF
1 83.4 94.3 82.5 98.8
2 82.1 96.1 81.8 98.5
3 86.5 94.9 82.1 98.8
4 82.3 95.3 81.7 99.2
5 84.8 96.5 82.3 98.9

Table 2. Recognition results of bronze inscriptions by different algorithms
View table
Table 2. Recognition results of bronze inscriptions by different algorithms
Algorithm Recall Precision F1 Accuracy /% Loss
LeNet5 0.431 0.422 0.438 43.45 2.55
Vgg13 0.788 0.778 0.811 81.9 1.06
Vgg16 0.793 0.816 0.79 81.75 1.27
RestNet 0.903 0.901 0.924 90.85 0.33
Proposed
algorithm
0.986 0.986 0.987 98.86 0.049

Tools

Get Citation

Copy Citation Text

Tongyao Luo, Huiqin Wang, Ke Wang, Zhan Wang, Hong Wang. Small Sample Bronze Inscription Classification Algorithm Based on Morphological Features[J]. Laser & Optoelectronics Progress, 2023, 60(4): 0410011

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites