No abstract

A number of vision-based methods for detecting laser-induced defects on optical components have been implemented to replace the time-consuming manual inspection. While deep-learning-based methods have achieved state-of-the-art performances in many visual recognition tasks, their success often hinges on the availability of a large number of labeled training sets. In this paper, we propose a surface defect detection method based on image segmentation with a U-shaped convolutional network (U-Net). The designed network was trained on paired sets of online and offline images of optics from a large laser facility. We show in our experimental evaluation that our approach can accurately locate laser-induced defects on the optics in real time. The main advantage of the proposed method is that the network can be trained end to end on small samples, without the requirement for manual labeling or manual feature extraction. The approach can be applied to the daily inspection and maintenance of optical components in large laser facilities.

No abstract

No abstract

With the progress in hardware implementation of artificial neural networks, the ability to analyze their faulty behavior has become increasingly important to their diagnosis, repair, reconfiguration, and reliable application. The behavior of feedforward neural networks with hardlimiting activation function under stuck-at faults is studied in this article. It is shown that the stuck-at-M faults have a larger effect on the network's performance than the mixed stuck-at faults, which in turn have a larger effect than that of stuck-at-0 faults. Furthermore, the fault-tolerant ability of the network decreases with the increase of its size for the same percentage of faulty interconnections. The results of our analysis are validated by Monte-Carlo simulations.