An improved cycle-consistent generative adversarial network(CycleGAN) method for defect data augmentation based on feature fusion and self attention residual module is proposed to address the insufficiency of defect s...An improved cycle-consistent generative adversarial network(CycleGAN) method for defect data augmentation based on feature fusion and self attention residual module is proposed to address the insufficiency of defect sample data for light guide plate(LGP) in production,as well as the problem of minor defects.Two optimizations are made to the generator of CycleGAN:fusion of low resolution features obtained from partial up-sampling and down-sampling with high-resolution features,combination of self attention mechanism with residual network structure to replace the original residual module.Qualitative and quantitative experiments were conducted to compare different data augmentation methods,and the results show that the defect images of the LGP generated by the improved network were more realistic,and the accuracy of the you only look once version 5(YOLOv5) detection network for the LGP was improved by 5.6%,proving the effectiveness and accuracy of the proposed method.展开更多
The efficiency balance phenomenon for see-through head-mounted displays with different microstructure con- ditions can be found both theoretically and using optical simulation software. A simple mathematical calculati...The efficiency balance phenomenon for see-through head-mounted displays with different microstructure con- ditions can be found both theoretically and using optical simulation software. A simple mathematical calculation is used to determine the relationship between the real image （see-through function） energy and the virtual image energy. The simulation is based on factors taken from previous research studies. It is found that the balance value of the optical efficiency remains almost constant （66.63% to 67.38%） under different microstructure conditions. In addition, suitable conditions for the microstructures in see-through head-mounted displays for daily applications can be predicted.展开更多
基金supported by the Jiangsu Province IUR Cooperation Project (No.BY2021258)the Wuxi Science and Technology Development Fund Project (No.G20212028)。
文摘An improved cycle-consistent generative adversarial network(CycleGAN) method for defect data augmentation based on feature fusion and self attention residual module is proposed to address the insufficiency of defect sample data for light guide plate(LGP) in production,as well as the problem of minor defects.Two optimizations are made to the generator of CycleGAN:fusion of low resolution features obtained from partial up-sampling and down-sampling with high-resolution features,combination of self attention mechanism with residual network structure to replace the original residual module.Qualitative and quantitative experiments were conducted to compare different data augmentation methods,and the results show that the defect images of the LGP generated by the improved network were more realistic,and the accuracy of the you only look once version 5(YOLOv5) detection network for the LGP was improved by 5.6%,proving the effectiveness and accuracy of the proposed method.
基金supported in part by the Ministry of Science and Technology,Taiwan,project number MOST104-2220-E-009-006in part by the "Aim for the Top University Plan" of the National Chiao Tung University and the Ministry of Education,Taiwan,China
文摘The efficiency balance phenomenon for see-through head-mounted displays with different microstructure con- ditions can be found both theoretically and using optical simulation software. A simple mathematical calculation is used to determine the relationship between the real image （see-through function） energy and the virtual image energy. The simulation is based on factors taken from previous research studies. It is found that the balance value of the optical efficiency remains almost constant （66.63% to 67.38%） under different microstructure conditions. In addition, suitable conditions for the microstructures in see-through head-mounted displays for daily applications can be predicted.
基金The National High Technology Research and Development Program major national special plan(No.2013AA030601-2)the National Natural Science Foundation of China(No.61405037)the Science and Technology Development Foundation for Fuzhou University(No.650074)
