Deep learning is a type of learning by Machine in which a model created learns to perform classification tasks directly from images in Vision. Deep learning is usually implemented using a neural network architecture with more number of layers in the network and hence called as deep learning. Unlike traditional neural networks contain only 2 or 3 layers, deep networks can have hundreds.
A deep neural network combines multiple nonlinear processing layers, using simple elements operating in parallel and inspired by biological nervous systems. It consists of an input layer, several hidden layers, and an output layer. The layers are interconnected via nodes, or neurons, with each hidden layer using the output of the previous layer as its input. Deep learning often requires hundreds of thousands or millions of images for the best results. It’s also computationally intensive and requires a high-performance GPU. Machine Vision based factory automation is one area that has embraced "Deep Learning" to play a role in classification, segmentation and recognition where normal conventional machine vision algorithms fail. Deeplearning-b ased algorithms provide higher inspection accuracy than that found in visual inspection and conventional machine vision inspection. With high-accuracy automated inspection capacity, it is possible for one person to manage more than one inspection equipment. By using GPU-processing languages, such as CUDA and CUDNN, deep learning algorithms with support for Multi-GPU and Multi-Threading, the image processing speed can be maximized. Not every application benefits from deep learning capabilities. Good candidates for deep learning then, may be inspection applications where there is no pre-defined shape. Deep learning and machine vision make a powerful combination. Deep learning enhances machine vision capabilities and opens up entirely new application possibilities. The two technologies are only just beginning to merge and have the potential to revolutionize the way machine vision is deployed in nearly every way.