Convolutional Neural Network (CNN)-Based Signature Verification via Cloud-Enabled Raspberry Pi System
Significant development in digital technology has contributed to the emergence of the Internet of Things (IoT). Machines can now interact between each other through the Internet without any human intervention. The rapid integration of the IoT has revolutionised many industries and expanded their cap...
Saved in:
Published in | Artificial Intelligence-based Internet of Things Systems pp. 191 - 217 |
---|---|
Main Authors | , , , |
Format | Book Chapter |
Language | English |
Published |
Cham
Springer International Publishing
|
Series | Internet of Things |
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | Significant development in digital technology has contributed to the emergence of the Internet of Things (IoT). Machines can now interact between each other through the Internet without any human intervention. The rapid integration of the IoT has revolutionised many industries and expanded their capacity to operate more autonomously and data-centrically. The presence of vast amount of interrelated computing device promotes the usage and capability of machine learning, which is the ability of a system to acquire its own knowledge by extracting patterns from raw data. Machine learning has been applied in many areas to increase productivity and security and enhance overall efficiency for the end user. One of them is to aid in verifying signatures to prevent fraudulent. A normal signature verification process is usually done electronically and visually, and it requires a tedious procedure. This is due to the high intrapersonal variability nature of a signature. To facilitate this process, we propose to integrate machine learning and the IoT in a portable scale to perform high accuracy verification system. This model uses a pre-trained convolutional neural network (CNN) on a Raspberry Pi. The CNN will analyse pixels from a signature image taken by the Pi camera to recognise abnormalities and differences and to identify false signature. Other than requiring a secure digital authentication to operate, it also informs the user immediately on the app execution and image being scanned via a cloud-based system. The system is expected to provide on-the-spot signature verification and minimise any logistic issues that stem from faulty signature to an organisation. |
---|---|
ISBN: | 3030870588 9783030870584 |
ISSN: | 2199-1073 2199-1081 |
DOI: | 10.1007/978-3-030-87059-1_7 |