Speckle Image Deformation Measurement Method Based on Convolutional Neural Network UNet++

Table 1. The training results of DIC-Net++
View table
View in Article
Table 1. The training results of DIC-Net++
AEE/pixel DIC-Net++-d DIC-Net++-s
Train set 0.033 2 0.009 8
Validation set 0.033 9 0.010 3

Table 2. Performance of DIC-Net++-d on self-built dataset

View table

View in Article

Table 2. Performance of DIC-Net++-d on self-built dataset

Group	Model	RMSE/pixel		MAE/pixel
Group	Model	u	v	u	v
0	StrainNet-f	0.036 54	0.035 85	0.027 66	0.026 89
	DIC-Net-d	0.025 23	0.025 66	0.019 25	0.019 28
	DIC-Net++-d	0.015 48	0.015 07	0.011 26	0.011 20
1	StrainNet-f	0.118 56	0.099 65	0.070 21	0.079 55
	DIC-Net-d	0.083 03	0.085 64	0.050 14	0.064 04
	DIC-Net++-d	0.016 17	0.016 00	0.012 38	0.012 19

Table 3. Performance of different models on DIC Challenge 1.0
View table
View in Article
Table 3. Performance of different models on DIC Challenge 1.0
Model RMSE/pixel
Sample 1 Sample 3 Sample 4
StrainNet-f 0.150 50 0.053 52 0.148 14
DIC-Net-d 0.236 44 0.020 89 0.106 48
DIC-Net++-d 0.067 84 0.018 28 0.085 58

Table 4. Performance of different models on Star5
View table
View in Article
Table 4. Performance of different models on Star5
Model SR σ^v MEI
SrainNet-f 285 0.022 6 6.441
DIC-Net-d 231 0.015 3 3.534
DIC-Net++-d 98 0.014 0 1.372

Table 5. Performance of different models on Star6
View table
View in Article
Table 5. Performance of different models on Star6
Model SR σ^v MEI
DIC-Net-s 82 0.009 936 0.008 1
DIC-Net++-s 124 0.005 459 2 0.003 7

Tools

Get Citation

Copy Citation Text

Qiang CHEN, Junzhe LIANG, Jin LIANG. Speckle Image Deformation Measurement Method Based on Convolutional Neural Network UNet++[J]. Acta Photonica Sinica, 2024, 53(11): 1112001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites