Computational Intelligence for Multimedia Big Data on the Cloud with Engineering Applications

Arun Kumar Sangaiah, Zhiyong Zhang, Michael Sheng

• Publisher: Academic Press
• Year: 2018
• Print ISBN: 9780128133149
• eBook ISBN: 9780128133279
• Topic: Technology & Engineering
• Subtopic: Automation in Engineering
• Index: Technology & Engineering

Computational Intelligence for Multimedia Big Data on the Cloud with Engineering Applications covers timely topics, including the neural network (NN), particle swarm optimization (PSO), evolutionary algorithm (GA), fuzzy sets (FS) and rough sets (RS), etc. Furthermore, the book highlights recent research on representative techniques to elaborate how a data-centric system formed a powerful platform for the processing of cloud hosted multimedia big data and how it could be analyzed, processed and characterized by CI. The book also provides a view on how techniques in CI can offer solutions in modeling, relationship pattern recognition, clustering and other problems in bioengineering. It is written for domain experts and developers who want to understand and explore the application of computational intelligence aspects (opportunities and challenges) for design and development of a data-centric system in the context of multimedia cloud, big data era and its related applications, such as smarter healthcare, homeland security, traffic control trading analysis and telecom, etc. Researchers and PhD students exploring the significance of data centric systems in the next paradigm of computing will find this book extremely useful. - Presents a brief overview of computational intelligence paradigms and its significant role in application domains - Illustrates the state-of-the-art and recent developments in the new theories and applications of CI approaches - Familiarizes the reader with computational intelligence concepts and technologies that are successfully used in the implementation of cloud-centric multimedia services in massive data processing - Provides new advances in the fields of CI for bio-engineering application

Chapter 1

A Cloud-Based Big Data System To Support Visually Impaired People

Huseyin Temucin^⁎; Ali Seydi Keceli^⁎; Aydin Kaya^⁎; Hamdi Yalin Yalic^⁎; Bedir Tekinerdogan^{† ⁎}Hacettepe University, Ankara, Turkey
^†Wageningen University, Wageningen, Netherlands

Abstract

In society, visual impairment is one of the important health issues that severely impede the daily life and welfare of many people. According to the 2014 World Health Organization (WHO) report, there exist 285 million visually impaired people worldwide, and more than 400 thousand in Turkey. To support the visually impaired people and likewise help them integrate into the society, several challenges need to be solved. In this study, we focus on two important issues, including the reading of normal, non-braille text, and face recognition. Reading of normal texts beyond Braille is one of the important life activities that is required in the daily personal and professional life of people. Face recognition is important for social interaction and communication. To solve both problems we propose a system which can help visually impaired people to recognize human faces and read normal text. The tool is based on a cloud-based architecture whereby services are provided for text and face recognition. The services are based on big data analytics together with deep learning algorithms. In this chapter, we discuss the overall architecture for such a text and face recognition system, the design decisions, the key challenges, the presented analytics approaches and the lessons learned that could be of value to both practitioners and researchers.

Keywords

Visually impaired; Face detection; Face recognition; Deep learning; Optical character recognition; Service oriented computing

1.1 Introduction

At the heart of people's ability to sustain their social lives is the acquisition of information and communication. Information acquisition is often achieved through reading, and communication is achieved through interaction with people. According to the 2014 World Health Organization (WHO) report [1], there exist 285 million visually impaired people worldwide, and more than 400 thousand in Turkey according to Turkish Statistics Institute [2]. It is known that the percentage of people with visual loss is increasing after the age of 40 years, which indicates that this is particularly a problem for the elderly people. Reading newspapers, books, and magazines, as well as the information they need to use in daily life (such as the telephone number of an institution) is an important problem for the visually impaired people.

Obviously, an increasing number of resources written in the Braille alphabet, together with an increasing number of audio books that also cover broader interest domains. Many national and local governments provide these services to disabled citizens. For example, at the National Library (Ankara, Turkey), voluntary readers record audio books on the demands of visually impaired members at the Bookstore and Visually Impaired Service Center [3]. National Library Service for the Blind and Physically Handicapped (NLS) provides Braille and audio books/magazines to those who needs [4]. Royal National Institute of Blind People (RNIB) provides access to more than 25,000 audio resources [5]. Despite these commercial and charitable services visually impaired people still have to cope with several different problems. Part of the problem is that learning a new alphabet such as Braille will take time and exercise before the visually impaired people can fully achieve a full information acquisition experience. Moreover, although the number of Braille and audio books are increasing still a large part of the information is not in these media.

In addition to reading, human interaction forms an important issue for the social communication and herewith the welfare of the visually impaired people. As known, human interaction is based on either or both voice and facial recognition. While visually impaired people can hear and recognize voices, they need support for facial recognition to enable the communication.

An important support for solving these problems is to provide auxiliary assistance devices, which will help visually impaired people in reading and human interaction. The devices to be designed will have to be portable devices that provide its user's functionality for reading and person recognition through ergonomic interfaces. Since modern text and facial recognition techniques are based on machine learning and require high computing power, mobile systems that provide high computational tasks like computer vision or text to speech conversion require strong processors and huge energy consumption. Although there are OCR and face recognition solutions for smartphones, none of these are applicable for daily scenarios. Commercial products have very high accuracy on scanned documents but for complicated images like those in real life, they have a high fault rate.

By the way, the assistive device to be designed needs to be ergonomic and low power consumption so that it can be used effectively in the outdoor. In this concept, the combination of IOT and cloud computing together will be an optimal solution for the mentioned scenario. In this scenario, IoT sensor will be used as an input device for data collection and those data will be processed in a service-oriented architecture.

In this paper, we report on an intelligent assistive device that implements the functionality for reading and face recognition. The device is based on the IoT architecture in which sensor data is connected to the cloud whereby information is derived using big data analytics and deep learning algorithms. The deep learning based OCR and Face recognition methods are running on a cloud service bucket. The proposed system will provide face detection and face recognition based on cloud services using face metadata on the cloud. The system also reads the texts aloud from properly printed documents, especially books, newspapers, and magazines on endpoint device after cloud process. While we have implemented a service based client layer, we have also implemented a portable client device prototype with the features described.

The remainder of the paper is organized as follows. Section 1.2 presents the related work. In Section 1.3 we provide the background that is necessary for understanding the provided solution. Section 1.4 presents the problem statement. Section 1.5 presents the system architecture and Section 1.6 discusses our machine learning approach including both the theoretical and implementation aspects. Section 1.7 presents the implemented prototype device. Section 1.8 presents the evaluation studies and finally, Section 1.9 presents the conclusion.

1.2 Related Work

Computational intelligence for cloud systems allows end users to access large scale and highly reliable computing systems. With the combination of big data and cloud computing, the vast amount of data could be analyzed without a high computational infrastructure. Arnaldo et al. [6] proposed a cloud-based machine learning framework. Their framework exploits data parallelism to allow its users to handle large data problems. Their framework also allows users to bring their own machine learning model in a plug and play style.

Gao ...

Table of contents

• Cover image
• Title page
• Table of Contents
• Copyright
• Contributors
• Foreword
• Preface
• Chapter 1: A Cloud-Based Big Data System To Support Visually Impaired People
• Chapter 2: Computational Intelligence in Smart Grid Environment
• Chapter 3: Patient Facial Emotion Recognition and Sentiment Analysis Using Secure Cloud With Hardware Acceleration
• Chapter 4: Novel Computational Intelligence Techniques for Automatic Pain Detection and Pain Intensity Level Estimation From Facial Expressions Using Distributed Computing for Big Data
• Chapter 5: Computational Intelligence Enabling the Development of Efficient Clinical Decision Support Systems: Case Study of Heart Failure
• Chapter 6: Aspect Oriented Modeling of Missing Data Imputation for Internet of Things (IoT) Based Healthcare Infrastructure
• Chapter 7: A Hybrid Computational Intelligence Decision Making Model for Multimedia Cloud Based Applications
• Chapter 8: Energy-Constrained Workflow Scheduling in Cloud Using E-DSOS Algorithm
• Chapter 9: Producing Better Disaster Management Plan in Post-Disaster Situation Using Social Media Mining
• Chapter 10: Metaheuristic Algorithms: A Comprehensive Review
• Chapter 11: Unsupervised Anomaly Detection for High Dimensional Data—an Exploratory Analysis
• Chapter 12: Fog – Driven Healthcare Framework for Security Analysis
• Chapter 13: Medical Quality of Service Optimization Over Internet of Multimedia Things
• Chapter 14: Energy-Efficiency of Tools and Applications on Internet
• Chapter 15: Transforming Healthcare Via Big Data Analytics
• Index