Highly Efficient Anomaly Detector for Industry Purposes

Product quality control is the leading trend in industrial production. It is heading toward the analysis of product before reaching the end customer. The most popular image analysis methods are based on deep neural networks. Their major advantages are fast performance, robustness, and the fact that they can be exploited even in complicated classification problems. The use of ML methods on high-performance production lines may be limited by inference time or training time. We propose our model for anomaly detection that needs very few training samples to reach required accuracy.