Strip Defect Classification Based on Improved Generative Adversarial Networks and MobileNetV3

Jiang Chang; Shengqi Guan; Hongyu Shi; Luping Hu; Yiqi Ni

doi:10.3788/LOP202158.0410016

Laser & Optoelectronics Progress, Volume. 58, Issue 4, 0410016(2021)

Strip Defect Classification Based on Improved Generative Adversarial Networks and MobileNetV3

Jiang Chang¹, Shengqi Guan^1,2、*, Hongyu Shi³, Luping Hu¹, and Yiqi Ni¹

Author Affiliations

¹School of Mechanical and Electronic Engineering, Xi'an Polytechnic University, Xi'an, Shaanxi 710048, China

²Shaoxing Keqiao West-Tex Textile Industry Innovative Institute, Shaoxing, Zhejiang 312030, China

³School of Computer Science, Xi'an Polytechnic University, Xi'an, Shaanxi 710048, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(6)

Fig. 1. Structure of the generator network

Download full size

Fig. 2. Structure of the discriminator network

Download full size

Fig. 3. Part of the original strip steel defect images

Download full size

Fig. 4. Images obtained by different network training. (a) 100 epochs; (b) 300 epochs; (c) 500 epochs

Download full size

Table 1. Structure of the improved MobileNetV3
View table
Table 1. Structure of the improved MobileNetV3
Operator Resolution Channel Layer
Conv 3×3 64×64 16 1
Bottleneck 3×3 32×32 24 2
Bottleneck 5×5 16×16 40 2
Bottleneck 5×5 8×8 80 2
Conv 1×1 & Pooling 4×4 120 1
Conv 1×1 1×1 240 1
Conv 1×1 1×1 6 1

Table 2. Image classification performance of different networks

View table

Table 2. Image classification performance of different networks

Network	Parameter amount /million	Predictedtime /ms	Accuracy /%	Precision /%	Recall /%	F₁
Original MobileNetV3	1.24	272	92.05	92.54	91.47	0.92
Improved MobileNetV3	0.09	231	94.67	95.21	94.79	0.95
VGG16	17.35	320	93.85	94.33	93.67	0.94
ResNet34	21.28	318	91.23	91.41	90.59	0.91
DenseNet121	6.95	315	94.36	94.50	94.10	0.94

Tools

Get Citation

Copy Citation Text

Jiang Chang, Shengqi Guan, Hongyu Shi, Luping Hu, Yiqi Ni. Strip Defect Classification Based on Improved Generative Adversarial Networks and MobileNetV3[J]. Laser & Optoelectronics Progress, 2021, 58(4): 0410016

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites