Convolutional Neural Networks

Convolutional neural networks (CNNs) are one of the most commonly used deep learning networks. They are feedforward artificial neural networks inspired by the animal's visual cortex. These networks are designed for data with spatial and temporal information. Therefore, convolutional neural networks are widely used in image and video recognition, speech recognition, and natural language processing. The architecture of convolution neural network consists of various layers which convert the raw input pixels into a class score.

For more details, see Learn About Convolutional Neural Networks.

You can train CNNs from scratch, by transfer learning, or by feature extraction. You can then use the trained network for classification or regression applications.

For more details on training CNNs, see Pretrained Deep Neural Networks.

For more details on deep learning, training process, and CNNs, see Deep Learning Onramp.