Matrox® Imaging recemtly released the Matrox Imaging Library (MIL) 10 Processing Pack 3 software update featuring a CPU-based, image classification module which makes use of deep learning technology for machine vision applications. Processing Pack 3 also includes the addition of a photometric stereo tool to bring out hard to spot surface anomalies and a new dedicated tool to locate rectangular features.
Deep learning for image classification
Leveraging convolutional neural network (CNN) technology, the Classification tool categorizes images of highly textured, naturally varying, and acceptably deformed goods. The inference is performed exclusively by Matrox Imaging-written code on a mainstream CPU, eliminating the dependence on third-party neural network libraries and the need for specialized GPU hardware. The intricate design and training of a neural network is carried out by Matrox Imaging, taking advantage of the accumulated experience, knowledge, and skill of its machine learning and machine vision experts.