Classification of Cultural Fragments Based on Adaptive Weights of Multi-Feature Descriptions

Zhengjie Lu; Chunhui Li; Guohua Geng; PengBo Zhou; Yan Li; Yang Liu

doi:10.3788/LOP57.041511

Laser & Optoelectronics Progress, Volume. 57, Issue 4, 041511(2020)

Classification of Cultural Fragments Based on Adaptive Weights of Multi-Feature Descriptions

Zhengjie Lu^1、**, Chunhui Li¹, Guohua Geng^1、*, PengBo Zhou², Yan Li¹, and Yang Liu¹

Author Affiliations

¹College of Information Science and Technology, Northwest University, Xi'an, Shaanxi 710127, China

²College of Arts and Media, Beijing Normal University, Beijing 100875, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(10)

Fig. 1. Flow chart of algorithm steps

Download full size

Fig. 2. Terracotta warriors fragment with distinctive features

Download full size

Fig. 3. Integral invariant

Download full size

Fig. 4. Feature computation. (a) G10-19-15(38) original fragment; (b) G10-19-15(38) local significant feature

Download full size

Fig. 5. G3-IV-14-5 lower body fragment

Download full size

Table 1. Fragment connectivity calculation result
View table
Table 1. Fragment connectivity calculation result
Fragment number Fragment screenshot Number of vertices Number of point sets
G10-19-15(38) 26543 49
G10-52-29 37932 73
G3-10-66 56281 89
G3-18-92 80208 173
G10-19-13 11868 23

Table 2. Classification result
View table
Table 2. Classification result
Fragmentation Amount Correctclassification Wrongclassification
Upper body 50 39 11
Lower body 57 39 18
Head 41 33 8
Hand 41 31 10
Legs and feet 42 30 12

Table 3. Classification accuracy
View table
Table 3. Classification accuracy
Fragmentation Accuracy /%
Upper body 78.00
Lower body 68.42
Head 80.48
Hand 75.00
Legs and feet 71.42

Table 4. Feature descriptor comparison
View table
Table 4. Feature descriptor comparison
Featuredescriptor Average describeduration /s Match timeper round /s
FPFH 399 183
SHOT 132 303
Descriptor ofthis paper 159 192

Table 5. FPFH classification accuracy
View table
Table 5. FPFH classification accuracy
Fragmentation Accuracy /%
Upper body 64.2
Lower body 53.5
Head 56.6
Hand 60.3
Legs and feet 59.6

Tools

Get Citation

Copy Citation Text

Zhengjie Lu, Chunhui Li, Guohua Geng, PengBo Zhou, Yan Li, Yang Liu. Classification of Cultural Fragments Based on Adaptive Weights of Multi-Feature Descriptions[J]. Laser & Optoelectronics Progress, 2020, 57(4): 041511

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites