Achieve 100% inspection rate with superhuman accuracy thanks to deep learning methods for visual inspection that are more robust than traditional machine vision systems to reduce quality control costs in defect inspection and material classification tasks.