Image Recognitions with Deep Learning and Comparison between Different Convolutional Neural Network Structures using Tensor flow

Abstract

Over the past decade, human movements and behaviors' can be monitored by using big data, including
Global Positioning System (GPS) data, social media login data and mobile phone tracking data. Deep Learning is the part
of an artificial intelligence function that imitates the workings of the human brain in processing data and creating
patterns for use in decision making. Deep learning is a subset of machine learning in Artificial Intelligence (AI) that it
capable of learning unsupervised from unstructured or unlabeled data. This is also called as Deep Neural Learning or
Deep Neural Network. The process of machine learning can be carried out with the help of hierarchical level of artificial
neural networks. Automatic number-plate recognition (ANPR) is a technology that uses optical character recognition on
images to create vehicle location data from vehicle number plate image. It is used to check if a vehicle is registered or
licensed. Automatic number plate recognition can be used to store the captured images by the cameras and also the text
from the license plate, with some configurable to store a photograph of the driver. However, traditional methods may not
be suitable for extracting comprehensive vehicle information due to the complexity and diversity of human behaviors.
Studies have shown that deep neural networks have out paced the abilities of human beings in various fields and that deep
neural networks can be explained in a unique manner. Thus, deep neural network methods can potentially be used to
understand human behaviors. In this project, a deep learning neural network constructed in Tensor Flow is applied to
detect and classify vehicle information in toll gates available across country and the models of these vehicles are analyzed
to verify the classification results. The vehicles doesn't need to spend lots of time on the toll gate, as the Tensor Flow
system will recognize the number plate and based on the government identification, the amount will auto debit from the
respective bank. For the social science classification problem investigated in this study, the deep neural network classifier
in Tensor Flow provides better accuracy and more lucid visualization than do traditional neural network methods, even
for erratic classification rules. Furthermore, the results of this study reveal that Tensor Flow has considerable potential
for application in the human geographyfield.