Convolutional Neural Network in Deep Learning for Object Tracking: A Review

In recent times, there is an enormous growth of data, which raises the need for a system that handles it efficiently. The emergence of deep learning and machine learning algorithms has improved the performance of automated image classification tasks. The Convolutional Neural Network (CNN) architecture is designed to process raw images with less manual preprocessing which makes it the most suitable model for various complex deep learning tasks. Every neural network architecture comprises smaller units called Neurons that perform a specific operation on the input data, leading to extraction of useful information called features. The learning capability of the Convolutional Neural Network is enhanced with the utilization of a training algorithm like Backpropagation and apart from these optimizers, there are several hyperparameters that can be fine-tuned in the system, without compromising the accuracy of the model. The research work explores the scope of CNN which will help researchers to have a broad comprehension of CNN.