Neural Networks are computing systems inspired by the biological neural networks that constitute the human brain. Convolutional Neural Networks (CNN) are a class of deep, feed-forward artificial neural networks, most commonly applied to analysing images. Deep Learning uses large CNNs to solve complex problems difficult or impossible to solve with so-called conventional computer vision algorithms. Deep Learning algorithms may be easier to use for some inspection/classification applications as they typically learn by example. They do not require the user to figure out how to classify or inspect parts.
Instead, in an initial training phase, they learn just by being shown many images of the parts to be inspected. After successful training, they can be used to classify parts, or detect and segment defects. This is powerful technology, offering solutions to applications previously impossible with traditional algorithmic approach. However, understanding when to use and when not to use it is very important, if you are to achieve the desired result. Multipix Imaging’s Technology Specialist are here to offer you experienced and knowledgeable advice on Deep Learning.
Only need to train images of good objects/scenes
Fast implementation
Training of good and defect objects/scenes are required.
Most Robust Results
Increasingly, it is at the Edge where the processing is taking place and this makes perfect sense as it is efficiently managing data to send it to the right place, make the right decision, at the right time. The processing is near the source of data, for example a smart camera, and is increasingly based on chipsets designed for AI/Deep Learning.
Graphics Processor Unit (GPU) is design specifically to rapidly process data. They perform parallel operations on multiple sets/sources of data and are increasingly used in AI/Deep Learning when repetitive processing is required which is specifically during the training phase.
The Industrial IoT, or Industry 4.0 is utilising edge computing to great effect and this includes machine vision devices for inspection and control of production processes. Deep Learning and vision is also being increasingly used in consumer focused products and not confined to the production-line.
Featured Products for Deep Learning
Deep Learning works by training a neural network, teaching it how to classify a set of reference images. The run-time performance of the process highly depends on how representative and extensive the set of reference images is. Using a single GPU typically accelerates the learning and the processing phases by a factor of 100