What is the convolutional neural network?

CNN is the first successful approach in deep learning for classifying data like images, videos, and speech.

Convolution is a mathematical operation where a function is applied to another function to result in a mixture of the two functions.

CNN works similarly to an Artificial neural network. A CNN arranges its neurons into a three-dimensional form, i.e., length, breadth, height. Convolution is good at detecting simple structures in an image and constructing more complex features. Deep learning uses CNN to recognize objects in an image, and CNN uses a feed-forward neural network whose connectivity pattern is inspired by the animal-visual cortex.

Layers

CNN has 3-main and two sub-layers:

• Input layer: accepts pixels of an image in the form of a matrix. The input layer of CNN receives 2-dimensional data as input and passes this input data to the corresponding layers.
• Hidden layer: carries out feature extraction by performing certain calculations and manipulation. There are multiple hidden layers that can perform feature extraction, they are:
  • Convolutional layer
  • Relu Layer
  • Pooling layer
• Output layer: a fully-connected layer that identifies an object in the image.

CNN is useful in applications like Image processing, Video recognition, and Natural Language processing.