Image Classification Using LeNet
Vaibhav Verdhan
Abstract
And you have taken that extraordinary step of learning deep learning. Deep learning is an evolving field. From the basic neural networks, we have now evolved to complex architectures solving a multitude of business problems. Deep neural network–powered image processing and computer vision capabilities are allowing us to create better cancer detection solutions, reduce pollution levels, implement surveillance systems, and improve consumer experience. Sometimes, a business problem demands a customized approach. We might design our own network to suit the business problem at hand and based on the image quality available to us. The network design will also consider the available computation power to train and execute the networks. The researchers across organizations and universities spent a good amount of time in collecting and curating the datasets, cleaning and analyzing them, designing architectures, training and testing them, and iterating to improve the performance. It takes a good amount of time and a lot of patience to make a path-breaking solution.