Fortifying Quality
Management with
Deep Learning
Why this use case matters
Deep learning, specifically computer vision and natural language processing, can be designed to identify defects during quality checks. These deep learning models can verify that products aren’t defective and labels/packaging present, correct, straight, and readable, saving huge amounts of repetitive work for increased process accuracy.
Techniques
Natural Language Processing, Computer Vision, Classification
Algorithms
Single-Shot Detection, YOLO, Mask-R CNN, Res-Net50, RanSaC, Filtering, Auto-Encoders, GPT-3, Transformers (RNNs), BERT, ERNIE, Word2vec
Outcome
Automating & improving human-based quality checks at scale.
Mosaic used deep learning to automate the work-intensive task of quality checking and rectifying information on product packaging.
Mosaic has compiled our industry expertise into a Machine Learning playbook for Manufacturing.
