An Investigation into Different Architectures of Convolutional Neural Networks

In Deep Learning, a Convolutional Neural Network (CNN) is used to study the various aspects of visual resources. CNN features can be used for advanced tasks like classification and segmentation of images, detection of objects and other complex tasks. The state-of-the-art models consist of stacked convolutional layers. But modern architectures construct convolutional layers by using new ideas. They allow CNN to work more efficiently. This paper analyses the performance of some important CNN architectures in particular applications.