Intelligent Technologies and

Applications First International

Conference INTAP 2018 Bahawalpur
Pakistan October 23 25 2018 Revised
Selected Papers Imran Sarwar Bajwa
Visit to download the full and correct content document:
https://textbookfull.com/product/intelligent-technologies-and-applications-first-internati
onal-conference-intap-2018-bahawalpur-pakistan-october-23-25-2018-revised-selecte
d-papers-imran-sarwar-bajwa/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Applications of Computing and Communication

Technologies: First International Conference, ICACCT
2018, Delhi, India, March 9, 2018, Revised Selected
Papers Ganesh Chandra Deka
https://textbookfull.com/product/applications-of-computing-and-
communication-technologies-first-international-conference-
icacct-2018-delhi-india-march-9-2018-revised-selected-papers-
ganesh-chandra-deka/

Software Technologies Applications and Foundations STAF

2018 Collocated Workshops Toulouse France June 25 29
2018 Revised Selected Papers Manuel Mazzara

https://textbookfull.com/product/software-technologies-
applications-and-foundations-staf-2018-collocated-workshops-
toulouse-france-june-25-29-2018-revised-selected-papers-manuel-
mazzara/

Computational Intelligence and Intelligent Systems 10th

International Symposium ISICA 2018 Jiujiang China
October 13 14 2018 Revised Selected Papers Hu Peng

https://textbookfull.com/product/computational-intelligence-and-
intelligent-systems-10th-international-symposium-
isica-2018-jiujiang-china-october-13-14-2018-revised-selected-
papers-hu-peng/

Simulation and Modeling Methodologies Technologies and

Applications 8th International Conference SIMULTECH
2018 Porto Portugal July 29 31 2018 Revised Selected
Papers Mohammad S. Obaidat
https://textbookfull.com/product/simulation-and-modeling-
methodologies-technologies-and-applications-8th-international-
conference-simultech-2018-porto-portugal-july-29-31-2018-revised-
Smart Cities Green Technologies and Intelligent
Transport Systems 7th International Conference
SMARTGREENS and 4th International Conference VEHITS
2018 Funchal Madeira Portugal March 16 18 2018 Revised
Selected Papers Brian Donnellan
https://textbookfull.com/product/smart-cities-green-technologies-
and-intelligent-transport-systems-7th-international-conference-
smartgreens-and-4th-international-conference-vehits-2018-funchal-
madeira-portugal-march-16-18-2018-revise/

Futuristic Trends in Network and Communication

Technologies: First International Conference, FTNCT
2018, Solan, India, February 9–10, 2018, Revised
Selected Papers Pradeep Kumar Singh
https://textbookfull.com/product/futuristic-trends-in-network-
and-communication-technologies-first-international-conference-
ftnct-2018-solan-india-february-9-10-2018-revised-selected-
papers-pradeep-kumar-singh/

Information and Communication Technologies for Ageing

Well and e-Health: 4th International Conference,
ICT4AWE 2018, Funchal, Madeira, Portugal, March 22–23,
2018, Revised Selected Papers Panagiotis D. Bamidis
https://textbookfull.com/product/information-and-communication-
technologies-for-ageing-well-and-e-health-4th-international-
conference-ict4awe-2018-funchal-madeira-portugal-
march-22-23-2018-revised-selected-papers-pana/

Web and Big Data APWeb WAIM 2018 International

Workshops MWDA BAH KGMA DMMOOC DS Macau China July 23
25 2018 Revised Selected Papers Leong Hou U

https://textbookfull.com/product/web-and-big-data-apweb-
waim-2018-international-workshops-mwda-bah-kgma-dmmooc-ds-macau-
china-july-23-25-2018-revised-selected-papers-leong-hou-u/

Geo informatics in Sustainable Ecosystem and Society

6th International Conference GSES 2018 Handan China
September 25 26 2018 Revised Selected Papers Yichun Xie

https://textbookfull.com/product/geo-informatics-in-sustainable-
ecosystem-and-society-6th-international-conference-
gses-2018-handan-china-september-25-26-2018-revised-selected-
Imran Sarwar Bajwa
Fairouz Kamareddine
Anna Costa (Eds.)

Communications in Computer and Information Science 932

Intelligent Technologies
and Applications
First International Conference, INTAP 2018
Bahawalpur, Pakistan, October 23–25, 2018
Revised Selected Papers

123
Communications
in Computer and Information Science 932
Commenced Publication in 2007
Founding and Former Series Editors:
Phoebe Chen, Alfredo Cuzzocrea, Xiaoyong Du, Orhun Kara, Ting Liu,
Krishna M. Sivalingam, Dominik Ślęzak, and Xiaokang Yang

Editorial Board
Simone Diniz Junqueira Barbosa
Pontifical Catholic University of Rio de Janeiro (PUC-Rio),
Rio de Janeiro, Brazil
Joaquim Filipe
Polytechnic Institute of Setúbal, Setúbal, Portugal
Ashish Ghosh
Indian Statistical Institute, Kolkata, India
Igor Kotenko
St. Petersburg Institute for Informatics and Automation of the Russian
Academy of Sciences, St. Petersburg, Russia
Takashi Washio
Osaka University, Osaka, Japan
Junsong Yuan
University at Buffalo, The State University of New York, Buffalo, USA
Lizhu Zhou
Tsinghua University, Beijing, China
More information about this series at http://www.springer.com/series/7899
Imran Sarwar Bajwa Fairouz Kamareddine
•

Anna Costa (Eds.)

Intelligent Technologies
and Applications
First International Conference, INTAP 2018
Bahawalpur, Pakistan, October 23–25, 2018
Revised Selected Papers

123
Editors
Imran Sarwar Bajwa Anna Costa
Department of Computer Science and IT Department of Computer Engineering
Islamia University of Bahawalpur and Digital Systems
Baghdad, Pakistan University of Sao Paulo
São Paulo, Brazil
Fairouz Kamareddine
Mathematical and Computer Sciences
Heriot-Watt University
Edinburgh, UK

ISSN 1865-0929 ISSN 1865-0937 (electronic)

Communications in Computer and Information Science
ISBN 978-981-13-6051-0 ISBN 978-981-13-6052-7 (eBook)
https://doi.org/10.1007/978-981-13-6052-7

Library of Congress Control Number: 2018967266

© Springer Nature Singapore Pte Ltd. 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
 Preface

The present book includes the papers accepted for the First International Conference on
Intelligent Technologies and Applications (INTAP 2018), held in Bahawalpur,
Pakistan, during October 23–25, 2018, organized by the Artificial Intelligence
Research Group with the collaboration of the Sir Sadiq Association of Computing and
hosted by the Islamia University of Bahawalpur. The conference was sponsored by the
Higher Education Commission, Pakistan.
The conference was organized in 13 simultaneous tracks: AI and Health (5), Sen-
timent Analysis (5), Intelligent Applications (7), Social Media Analytics (6), Business
Intelligence (6), Natural Language Processing (5), Information Extraction (4), Machine
Learning (6), Smart Systems (5), Semantic Web (6), Decision Support Systems (6),
Image Analysis (7), and Automated Software Engineering (6).
We received 251 submissions, from 28 countries and districts representing all
continents. After a blind review process, only 67 were accepted as full papers and
seven were selected as short papers based on the classifications provided by the Pro-
gram Committee, resulting in an acceptance rate of 29%. The selected papers reflect
state-of-the-art research work in the different domains and applications of artificial
intelligence and highlight the benefits of intelligent and smart systems in various fields
of life. These high-quality standards will be maintained and reinforced at INTAP 2019,
to be held at Harriot-Watt University, Scotland, and in future editions of this
conference.
Furthermore, INTAP 2018 included four plenary keynote lectures given by Letizia
Jaccheri (Norwegian University of Science and Technology, Norway), Julia Sidorova
(Blekinge Institute of Technology, Sweden), M. Abbas Choudhary (DIHE, Karachi),
Irfan Hyder (IoBM, Karachi), Dr. Riaz ul Amin (BUITEMS, Quetta), and Dr. Khurram
Khurshid (IST, Islamabad). We would like to express our appreciation to all of them
and in particular to those who took the time to contribute with a paper to this book.
On behalf of the conference Organizing Committee, we would like to thank all
participants. First of all, the authors, whose quality work is the essence of the con-
ference., and the members of the Program Committee, who helped us with their
expertise in reviewing and selecting the quality papers for this book. It is well known
that organizing an international conference requires the effort of many individuals. We
wish to thank also all the members of our Organizing Committee, whose work and
commitment were invaluable.

October 2018 Anna Costa

Imran Sarwar Bajwa
Fairouz Kamareddine
Julia Sidorova
 Organization

General Co-chairs
Imran Sarwar Bajwa The Islamia University of Bahawalpur, Pakistan
Mark G. Lee University of Birmingham, UK
Anna Helena Reali Costa University of São Paulo, Brazil

Program Co-chairs
Fairouz Kamareddine Heriot-Watt University, UK
Imran Ghani Indiana University of Pennsylvania, USA
Jamal Bentahar Concordia University, Canada
Dayang Norhayati A. Universiti Teknologi Malaysia, Malaysia
Jawawi

Organizing Committee
Amir Hussain University of Stirling, UK
Irfan Hyder Institute of Business Management, Pakistan
Omair Shafiq Carleton University, Canada
M. Abbas Choudhary Dadabhoy Institute of Higher Education, Karachi,
Pakistan
Noreen Jamil Unitec Institute of Technology, New Zealand
Ghulam Alli Mallah SALU, Khairpur, Pakistan
Riaz ul Amin BUITEMS University, Quetta, Pakistan
Aman Ullah Yasin CASE, Islamabad, Pakistan
Imran Memon Zhejiang University, China
Rafaqut Hussain Kazmi The Islamia University of Bahawalpur, Pakistan

Program Committee
Adel Al-Jumaily University of Technology Sydney, Australia
Adina Florea University Politehnica of Bucharest, Romania
Adriano V. Werhli Universidade Federal do Rio Grande, Brazil
Agostino Poggim Università degli Studi di Parma, Italy
Ales Zamuda University of Maribor, Slovenia
Alexander Gelbukh National Polytechnic Institute, Mexico
Amin Beheshti Macquarie University, Australia
Anand Nayyar Duy Tan University, Vietnam
António Luís Lopes Instituto Universitário de Lisboa, Portugal
Anna Helena Reali Costa University of São Paulo, Brazil
Alvaro Rubio-Largo Universidade NOVA de Lisboa, Portugal
VIII Organization

Asif Baba, Tuskegee University of Alabama, USA

Auxiliar Pedro Quaresma University of Coimbra, Portugal
Barbara Ongaro Liceo Alessandro Greppi, Milan, Italy
Bahram Amini Foulad Institute of Technology, Malaysia
Bernard Moulin Université Laval, Canada
Bujor Pavaloiu University Politehnica of Bucharest, Romania
Carl James Debono University of Malta, Malta
Carlos Filipe da Silva University of Minho, Portugal
Portela
Costin Badica University of Craiova, Romania
Chrisa Tsinaraki Technical University of Crete, Greece
Cyril de Runz Université de Reims Champagne-Ardenne, France
Dan Cristea UAIC, Romania
Di Wu North Dakota State University, USA
Dion Goh Hoe Lian Nanyang Technological University, Singapore
Elias Kyriakides KIOS Research Center, Cyprus
Eric Matson Purdue University, USA
Emanuele Principi Università Politecnica delle Marche, Italy
Farshad Fotouhi Wayne State University, USA
Francesca Alessandra LISI Università degli Studi di Bari, Italy
Gazi Erkan Bostanci Ankara University, Turkey
Gerald Schaefer Loughborough University, UK
Gianluca Reali University of Perugia, Italy
Gianluigi Ferrari Università degli studi di Parma, Italy
Giuseppe Boccignone University of Milan, Italy
Grigore Stamatescu Politehnica University of Bucharest, Romania
Hichem Omrani CEPS/INSTEAD, Luxembourg
Harald Kosch University of Passau, Germany
Haralambos Mouratidis University of Brighton, UK
Hazart Ali COMSATS Institute of Information Technology,
Abbottabad
Icsabel De La Torre Díez University of Valladolid, Spain
Imran Memon Zhejiang University, China
Jan Platos VŠB-TU Ostrava, Czech Republic
Jan Muhammad BUITEMS, Quetta, Pakistan
Jamal Bentahar Concordia University, USA
José Carlos Martins Fonseca University of Coimbra, Portugal
José Moreira Universidade de Aveiro, Portugal
José Torres Universidade Fernando Pessoa, Portugal
Juan Carlos Nieves Umeå Universitet, Sweden
Juha Röning University of Oulu, Finland
Jurek Z. Sasiadek Carleton University, Canada
Luis Álvarez Sabucedo Universidade de Vigo, Spain
Luis Fernandez Luque Salumedia, Seville, Spain
Luis Iribarne University of Almería, Spain
Luis Jimenez Linares Escuela Superior de Informática, Spain
 Organization IX

Luis Rodríguez Benítez Universidad de Castilla-la Mancha, Spain

Mariachiara Puviani Università di Modena e Reggio Emilia, Italy
Marko Hölbl University of Maribor, Slovenia
Maxime Morge Université de Lille, France
M. R. Spruit Universiteit Utrecht, The Netherlands
M. Asif Naeem Auckland University of Technology, New Zealand
M. Shamsul Islam Edith Cowan University, Australia
Marcin Pietron AGH, University in Kraków, Poland
Marjan Mernik University of Maribor, Slovenia
Monireh Ebrahimi Wright State University Ohio, USA
Muhammad Taimoor Khan RISC Software GmbH, Austria
Natalia Bilici Université du Luxembourg, Luxembourg
Noreen Jamil Unitec Institute of Technology, New Zealand
Omair Shafiq Carleton University, Canada
Paulo Urbano Universidade de Lisboa, Portugal
Preben Hansen The Swedish Institute of Computer Science, Sweden
Ramoni Lasisi Virginia Military Institute, USA
Raymond Wong The University of New South Wales, Australia
Ravi Jhawar SaToSS, Université du Luxembourg, Luxembourg
Riaz-ul-Amin BUITEMS, Quetta, Pakistan
Ricardo Campos Instituto Politécnico de Tomar, Portugal
Rodríguez García Daniel Autonomous University of Barcelona, Spain
Roslina Binti Salleh Universiti Teknologi Malaysia, Malaysia
Rung Ching Chen Chaoyang University of Technology, Taiwan
Ryszard Tadeusiewicz AGH University of Science and Technology, Poland
Roland Traunmüller University of Linz, Austria
Ruggero Donida Labati Università degli Studi di Milano, Italy
Samir B. Belhaouri University VT, Saudi Arabia
Smaranda Belciug University of Craiova, Romania
Soheila Abrishami Florida State University, USA
Stefan Schulz Medical University of Graz, Austria
Stefka Stoyanova Fidanova Bulgarian Academy of Sciences, Bulgaria
Tatjana Sibalija Belgrade Metropolitan University, Serbia
Thepchai Supnithi Sirindhorn International Institute of Technology,
Thailand
Thierry Badard Université Laval, Canada
Tomislav Stipancic FMENA Zagreb, Croatia
Václav Snášel Technical University of Ostrava, Czech Republic
Vilem Novak University of Ostrava, Czech Republic
Vladimir Filipović University of Belgrade, Serbia
Weronika T. Adrian University of Calabria, Italy
Wie Wie Xi’an University of Technology, China
William Bill Grosky University of Michigan-Dearborn, USA
Yap Bee Wah Universiti Teknologi MARA, Malaysia
Yasushi Kambayashi Nippon Institute of Technology, Japan
Zbynek Raida Brno University of Technology, Czech Republic
X Organization

Invited Speakers
Letizia Jaccheri Norwegian University of Science and Technology,
Norway
Julia Sidorova Blekinge Institute of Technology, Sweden
M. Abbas Choudhary Dadabhoy Institute of Higher Education, Karachi,
Pakistan
Syed Irfan Hyder Institute of Business Management, Karachi, Pakistan
 Contents

AI and Health

Enhanced Medical Image De-noising Using Auto Encoders and MLP . . . . . . 3

Seshadri Sastry Kunapuli, Praveen Chakravarthy Bh,
and Upasana Singh

E-BRACE: A Secure Electronic Health Record Access Method

in Medical Emergency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Shraddha Nayak, Md. Akbar Hossain, Farhaan Mirza, M. Asif Naeem,
and Noreen Jamil

Enhanced Fuzzy Resolution Appliance for Identification

of Heart Disease in Teenagers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Arfa Hassan, H. M. Bilal, M. Adnan Khan, M. Farhan Khan,
Rubina Hassan, and M. Sajid Farooq

Similarity Measuring for Clustering Patient’s Reports in Telemedicine . . . . . . 38

Ateya Iram and Sajid Habib Gill

A Review of Machine Learning for Healthcare Informatics

Specifically Tuberculosis Disease Diagnostics. . . . . . . . . . . . . . . . . . . . . . . 50
Priyanka Karmani, Aftab Ahmed Chandio, Imtiaz Ali Korejo,
and Muhammad Saleem Chandio

Sentiment Analysis

Long-Term Trends in Public Sentiment in Indian Demonetisation Policy . . . . 65

Adi Darliansyah, Herman Masindano Wandabwa, M. Asif Naeem,
Farhaan Mirza, and Russel Pears

Sentiment Analysis on Automobile Brands Using Twitter Data . . . . . . . . . . . 76

Zain Asghar, Tahir Ali, Imran Ahmad, Sridevi Tharanidharan,
Shamim Kamal Abdul Nazar, and Shahid Kamal

Sentiment Analysis of Student’s Facebook Posts . . . . . . . . . . . . . . . . . . . . . 86

Ateya Iram

Opinion and Emotion Mining for Pakistan General Election 2018

on Twitter Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Suleman Khan, Syed Atif Moqurrab, Rotaba Sehar, and Umair Ayub
XII Contents

Sentimental Analysis of Social Media to Find Out Customer Opinion . . . . . . 110

Haq Nawaz, Tahir Ali, Ali Al-laith, Imran Ahmad,
Sridevi Tharanidharan, and Shamim Kamal Abdul Nazar

Intelligent Applications

Parsing RDFs to Extract Object Oriented Model Using Apache Jena . . . . . . . 119
Umar Farooq Shafi, Hina Sattar, Imran Sarwar Bajwa, and Amna Ikram

Principle Features of Beamforming and Phase Shift of Phased

Array Antennas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Muhammad Saleem, Sidra Naz, and Anila Kauser

Generating Linked Data Repositories Using UML Artifacts . . . . . . . . . . . . . 142

Aqsa Khan and Saleem Malik

Containers vs Virtual Machines for Auto-scaling Multi-tier

Applications Under Dynamically Increasing Workloads . . . . . . . . . . . . . . . . 153
Muhammad Abdullah, Waheed Iqbal, and Faisal Bukhari

Anti-phishing Models for Mobile Application Development:

A Review Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Javaria Khalid, Rabiya Jalil, Myda Khalid, Maliha Maryam,
Muhammad Aatif Shafique, and Wajid Rasheed

Sensing Time Optimization Using Genetic Algorithm in Cognitive

Radio Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Muhammad Nadeem Ali, Iqra Naveed, Muhammad Adnan Khan,
Ayesha Nasir, and M. Tahir Mushtaq

A Bio-Inspired Rooted Tree Algorithm for Optimal Coordination

of Overcurrent Relays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Abdul Wadood, Tahir Khurshaid, Saeid Gholami Farkoush,
Chang-Hwan Kim, and Sang-Bong Rhee

Social Media Analytics

Social Media Competitive Analysis of Shoe Brands

on Customer Experiences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Imran Ahmad, Tahir Ali, Asad Nazir, and Shahid Kamal

A Fuzzy Logic Model for Evaluating Customer Loyalty in e-Commerce . . . . 216

Aiman Ashfaq and Mobeen Kausar

Tweets Competitive Sentimental Analysis of Android Mobile Brands

to Understand Customer Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
Umair Liaquat Ali, Tahir Ali, Imran Ahmad, and Shahid Kamal
 Contents XIII

Counter Terrorism on Online Social Networks Using

Web Mining Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Fawad Ali, Farhan Hassan Khan, Saba Bashir, and Uzair Ahmad

Analysis of Twitter Usage in Educational Institutions of Pakistan . . . . . . . . . 251

Gul Mina, Bakhtiar Kasi, Abdul Samad, and Riaz UlAmin

Detecting Suspicious Discussion on Online Forums Using Data Mining. . . . . 262

Haroon ur Rasheed, Farhan Hassan Khan, Saba Bashir,
and Irsa Fatima

Business Intelligence

Process Model Abstraction: Identifying Business Significant Activities . . . . . 277

Basharat Fatima and Khurram Shahzad

Automated Consistency Management in BPMN Based Business

Process Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Mamoona Ishaq and M. Abbas Choudhary

The Use of Fuzzy Logic in Creating a Visual Data Summary

of a Telecom Operator’s Customer Base . . . . . . . . . . . . . . . . . . . . . . . . . . 301
Julia Sidorova, Lars Sköld, Håkan Lennerstad, and Lars Lundberg

Social Media Competitive Analysis - A Case Study in the Pizza Industry

of Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Muhammad Usama Nazir, Sridevi Tharanidharan, M. Saleem Mian,
Imran Ahmad, Khizer Hayat, Shamim Kamal Abdul Nazar,
Shanza Zaman, Sohail Mustafa, and Muhammad Rehan Ghumman

Diffusion of Big Open Data Policy Innovation in Government

and Public Bodies in Pakistan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
Muhammad Mahboob Khurshid, Nor Hidayati Zakaria, Ammar Rashid,
Rafaqat Kazmi, and Muhammad Nouman Shafique

Transaction and Identity Authentication Security Model for E-Banking:

Confluence of Quantum Cryptography and AI . . . . . . . . . . . . . . . . . . . . . . 338
Tayyabah Hassan and Fahad Ahmed

Natural Language Processing

Tagging Assistant for Scientific Articles. . . . . . . . . . . . . . . . . . . . . . . . . . . 351

Zara Nasar, Syed Waqar Jaffry, and Muhammad Kamran Malik

A Natural Language Based Approach to Generate Document Stores . . . . . . . 363

Tayyaba Sana and Omair Shafiq
XIV Contents

Parallel String Matching for Urdu Language Text . . . . . . . . . . . . . . . . . . . . 369

Mirza Baber Baig and Taoshen S. Li

Generating SBVR-XML Representation of a Controlled Natural Language . . . 379

Shafaq Arshad, Imran Sarwar Bajwa, and Rafaqut Kazmi

Natural Language Based SQL Query Verification Against

Relational Schema. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
Shoaib Saleem Khan, Abid Saeed, Yasir Majeed,
and Muhammad Kamran

Information Extraction

Information Extraction of Ecological Canal System Based

on UAV Remote Sensing Data for Precision Irrigation. . . . . . . . . . . . . . . . . 403
Zichao Zhang, Yu Han, Jian Chen, Shubo Wang, Nannan Du,
Guangqi Wang, and Yongjun Zheng

Vehicle Detection, Tracking and Counting on M4 Motorway Pakistan . . . . . . 414

Ayesha Ansari, Khan Bahadar Khan, Muhammad Moin Akhtar,
and Hammad Younis

Legal Data Mining from Civil Judgments. . . . . . . . . . . . . . . . . . . . . . . . . . 426

Shahmin Sharafat, Zara Nasar, and Syed Waqar Jaffry

A Graph Theory Based Method to Extract Social Structure in the Society . . . 437
Wajid Rafique, Maqbool Khan, Nadeem Sarwar, Muhammad Sohail,
and Asma Irshad

Machine Learning

A Deep-Learning-Based Low-Altitude Remote Sensing Algorithm

for Weed Classification in Ecological Irrigation Area. . . . . . . . . . . . . . . . . . 451
Shubo Wang, Yu Han, Jian Chen, Yue Pan, Yi Cao, Hao Meng,
and Yongjun Zheng

Repairing Broken Links Using Naive Bayes Classifier . . . . . . . . . . . . . . . . . 461

Faheem Nawaz Khan, Adnan Ali, Imtiaz Hussain, Nadeem Sarwar,
and Hamaad Rafique

Predicting Web Vulnerabilities in Web Applications

Based on Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
Muhammad Noman Khalid, Humera Farooq, Muhammad Iqbal,
Muhammad Talha Alam, and Kamran Rasheed
 Contents XV

Malwares Detection for Android and Windows System by Using Machine

Learning and Data Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485
Syed Fakhar Bilal, Saba Bashir, Farhan Hassan Khan,
and Haroon Rasheed

Machine Learning Based Fault Diagnosis in HVDC Transmission Lines . . . . 496

Raheel Muzzammel

Smart Systems

Plant Irrigation and Recommender System–IoT Based Digital Solution

for Home Garden . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513
Sehrish Munawar Cheema, Museb Khalid, Abdur Rehman,
and Nadeem Sarwar

Smart Road-Lights and Auto Traffic-Signal Controller

with Emergency Override. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
Mohammad Faisal Naseer, Khan Bahadar Khan,
Muhammad Sannan Khaliq, and Muhammad Raheel

An Ecological Irrigation Canal Extraction Algorithm

Based on Airborne Lidar Point Cloud Data . . . . . . . . . . . . . . . . . . . . . . . . 538
Guangqi Wang, Yu Han, Jian Chen, Yue Pan, Yi Cao, Hao Meng,
Nannan Du, and Yongjun Zheng

Machine Learning for Analyzing Gait in Parkinson’s Patients

Using Wearable Force Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548
Asma Channa, Rahime Ceylan, and Attiya Baqai

Microchip with Advance Human Monitoring Technique and RFTS . . . . . . . . 560

Nadeem Sarwar, Faheem Nawaz Khan, Adnan Ali, Hamaad Rafique,
Imtiaz Hussain, and Asma Irshad

Semantic Web

Enrich Exiting UML Class Model Using Ontology . . . . . . . . . . . . . . . . . . . 573

Maria Iqbal and Abdur Rehman

OntoGen Based Ontology Concepts Generation from Graph . . . . . . . . . . . . . 579

Abid Saeed, Muhammad Shahzad Kamran, Shoaib Saleem Khan,
and Rao Muhammad Kamran

An Automatable Approach for Triples to PROV-O Mapping . . . . . . . . . . . . 591

Ayesha Mehmood, Amna Mehmood, and Bakhtawer Akhtar
XVI Contents

Relational Database to Resource Description Framework and Its Schema . . . . 604

Muhammad Faheem, Hina Sattar, Imran Sarwar Bajwa,
and Wasif Akbar

Generating RDFS Based Knowledge Graph from SBVR . . . . . . . . . . . . . . . 618

Bakhtawer Akhtar, Ayesha Mehmood, Amna Mehmood,
and Waheed Noor

Enrich Existing Ontologies with New Knowledge from Existing Artifacts . . . 630
Amna Mehmood, Ayesha Mehmood, and Bakhtawer Akhtar

Decision Support Systems

Decision Support System for Visualization of Tree Plantation

in Upper Sindh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645
Jamil Ahmed Chandio, Ghulam Ali Malah, Ubaidullah alias Kashif,
Yasir Ali Solangi, and Aadil Jameel

A Novel Approach to Generate OWL2 Models Using Case

Based Reasoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 656
Faiza Ali and M. Abbas Choudhary

Emergency Feedback System Based on SSVEP Brain

Computing Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
Tarwan Kumar Khatri, Humera Farooq, Muhammad Talha Alam,
Muhammad Noman Khalid, and Kamran Rasheed

Connection Time for Routing Decisions in Vehicular Delay

Tolerant Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 679
Adnan Ali, Nadeem Sarwar, Hamaad Rafique, Imtiaz Hussain,
and Faheem Nawaz Khan

Fire Controller System Using Fuzzy Logic for Safety . . . . . . . . . . . . . . . . . 691

Mobeen Kausar, Barera Sarwar, and Aimen Ashfaq

An Efficient Clustering of Wireless Sensor Network by Spectral

Graph Partitioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 698
Sonia Salman and Husnain Mansoor Ali

Image Analysis

Digital Image Steganography by Using a Hash Based LSB

(3-2-3) Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 713
Imra Aqeel and Muhammad Raheel
 Contents XVII

An Enhancement Method of Obstacle Information Obtaining Accuracy

in Binocular Vision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725
Zichao Zhang, Yu Han, Jian Chen, Wenhao Dou, Shubo Wang,
Nannan Du, Guangqi Wang, and Yongjun Zheng

Comparative Analysis of Pigment Network as a Feature

for Melanoma Detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 735
Umair Shafiq, Uzma Jamil, and Nafees Ayub

A Performance Assessment of Rose Plant Classification

Using Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745
Muzamil Malik, Amna Ikram, Syeda Naila Batool, and Waqar Aslam

Artefacts Removal from EEG Recordings in Urban Environment . . . . . . . . . 757

Muhammad Talha Alam, Humera Farooq, Muhammad Noman Khalid,
Tarwan Kumar, and Kamran Rasheed

A Survey on Digital Image Steganography Approaches . . . . . . . . . . . . . . . . 769

Imra Aqeel and Muhammad Babar Suleman

A Deep Neural Network Approach for Classification of Watermarked

and Non-watermarked Images. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 779
S. S. Tirumala, Noreen Jamil, and M. G. Abbas Malik

Automated Software Engineering

Multi-agent System Using Scrum Methodology for Software

Process Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 787
Shanawar Ali, Hafiz Hassan Ali, Sakha Qayyum, Fatima Sohail,
Faiza Tahir, Sahar Maqsood, and Mahum Adil

Generating a Referent Graph from Semantics of Business Vocabulary

and Business Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793
Muhammad Shahzad Kamran, Abid Saeed, and Abdul Hameed

Requirement Elicitation for Bespoke Software Development:

A Review Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805
Rabiya Jalil, Javaria Khalid, Maliha Maryam, Myda Khalid,
Sadaf Nawaz Cheema, and Iqra Iqbal

Quantitative Based Mechanism for Resolving Goals Conflicts

in Goal Oriented Requirement Engineering. . . . . . . . . . . . . . . . . . . . . . . . . 822
Taimoor Hassan, Muhammad Zunnurain Hussain,
Muhammad Zulkifl Hasan, Zaka Ullah, and Noor-ul Qamar
XVIII Contents

Automated Verification of Software Constraints Using Business Rules . . . . . 832

Sidra Sabir and Munsub Ali

Process Model Matching with Word Embeddings . . . . . . . . . . . . . . . . . . . . 838

Khurram Shahzad, Safia Kanwal, and Kamran Malik

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 851

AI and Health
 Enhanced Medical Image De-noising Using
Auto Encoders and MLP

Seshadri Sastry Kunapuli(&), Praveen Chakravarthy Bh(&),

and Upasana Singh(&)

Xvidia Technologies, Gurugram 122001, Haryana, India

{seshadri,praveen,upasana}@xvidia.net

Abstract. Preserving the original characteristics of an image which is trans-

mitted across a channel having different kinds of noise (i.e., either, uniform,
linear or Gaussian noise) is a crucial task, hence it has become a state of art for
the researchers in retrieving the original characteristics of the image by using
different denoising and image retrieving techniques. In earlier, many techniques
have been proposed such as patch wise denoising (e.g., Sliding Window), block
matching (e.g., BM3D), shallow and wide deep learning algorithms which
achieved a promising accuracy, yet failing in preserving the prominent char-
acteristics of an image which is a crucial task in Bio-Medical Instrumentation
systems. So, we proposed few algorithms which could preserve the smallest
possibilities of denoising the medical images and achieved a maximum accuracy
of 99.98% for SDAE (In Tensorflow Background), 99.97% for SDAE (In
Theano Background) and 99.99% for Multi-Layer Perception (MLP) technique
and later compared these with the accuracies of the existing methods.

Keywords: SDAE
MLP
Medical image
De-noising

1 Introduction

Digital Images extracted from the Bio-Medical Instrumentation system consists of

medical images plus noise (noise which was occurred due to non-linearity of the
devices or from the image acquisition techniques). In the field of Bio-Medical Engi-
neering a small amount of error may cause a lot more damage as the readings are taken
from a person whose characteristics varies accordingly (Black Box). Hence we need a
lot more improvement in enhancing the quality of the image thus preserving the
original characteristics by using different kinds of denoising techniques starting from
wavelet domain to spatial domain, patch based to block matching 3D fusion and other
shallow and Deep learning methods.
According to a survey [1] on various image denoising techniques, a basic way of
denoising is done in two approaches, i.e., Spatial Domain and Transform Domain.
Which intern divided into different sub-categories based on type of filter, transforma-
tion is used. In general, few Machine Learning algorithms which are commonly used in
Image Denoising are Filters, Wavelet Transforms and low pass filter with Fast Fourier
transforms and Singular Wavelet Transforms. The Metrics used in finding the accuracy

© Springer Nature Singapore Pte Ltd. 2019

I. S. Bajwa et al. (Eds.): INTAP 2018, CCIS 932, pp. 3–15, 2019.
https://doi.org/10.1007/978-981-13-6052-7_1
4 S. S. Kunapuli et al.

is similar in all the Machine learning algorithms i.e., Mean Square Error (MSE), PSNR
(Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index Metric).
Dabov [2] et al., proposed an image denoising algorithm using sparse transfor-
mation of 2D image fragments into 3D data arrays resulting a group of jointly filtered
image blocks. The collaborative filtering technique consists of three successive steps
such as, 3D Transformation of a group, shrinkage of a spectrum and Inverse-3D
transformation which results a successive image denoising technique. The metric used
in his 3D fusion technique is Peak Signal to Noise Ratio (PSNR) and Perspective
visualization. He has used different types of wavelet transformation techniques such as
DCT, DST and Walsh-Hadmard Transforms with different wavelets (Haar, Bior, Sym
etc.) into consideration with the BM3D Denoising technique.
An another algorithm improving the performance of denoising is done by imple-
menting Non-Local Means (NLM) and Block Matching 3D (BM3D) Fusion technique
has been proposed by Talebi [3] et al. by designing a global filter which can be applied
to the images to enhance the patch based methods by estimating a pixel from all the
pixels of the image. He stated that the denoising approaches such as bilateral filter,
LARK, BM3D, NLMS and (Patch Based Locally optimum filter) PLOW are data
dependent filtering schemes which uses each pixel individually from the neighbouring
pixels. In his work, he used Nystroms extension for image segmentation and decom-
posed them into corresponding Eigen vectors and applied them to an iterative and
trucking filter (e.g., BM3D and NLMS) and calculated PSNR for those images which
were later compared with his global filtered techniques.
Many Shallow Deep learning and Machine learning algorithms has come across
after that, but the real question is that, did the deep learning algorithms can keep up an
accuracy with the existing models? This question was later on explained by Burger [4]
et al., in his proposed paper. In his work, he applied a simple Multi-Layer Perception
(MLP) algorithm to the image patches and observed the resultant metrics (PSNR) for
different types of Machine Learning techniques such as GSM, KSVD and BM3D
techniques at different Noise levels. But the main disadvantage is that he compiled that
model using GPU instead of CPU which would be very difficult in the real time
scenarios as only few men can afford them.
Image Denoising of Ultra Low dose CT scan images were done by Nishio [5] et al.
In his work, he investigated the performance of Patch based Neural Network trained for
Ultra Low DCT images using Convolution Auto Encoders (CAE) and compared its
performance with NLMS & BM3D techniques using PSNR.
SSIM. As DAE ignores 2D image structure and CAE preserves the 2D image
structural data he used CAE instead of DAE for denoising medical Image datasets.
Denoising Auto Encoders using CNN architecture was designed by Gondara [6]
et al. In his work, he explained that the performance of image denoising can be further
enhanced by exploiting strong spatial correlations by constructing DAEs with Con-
volution layers. He tested his architecture on mini-MIAS database and Dental Radio
Graph database and achieved a strong SSIM of 0.89 and 0.90 when comparing to
NLMS and Median Filters.
So In this paper, we proposed a new Tensorflow and Theano architectures for DAE
which is having only three Dense layers and achieved a maximum accuracy of 99.98%
and 99.97% with Visual Proximities and compared with the novel architecture of our
 Enhanced Medical Image De-noising Using Auto Encoders and MLP 5

Multi Linear Perception (MLP) Neural Network which achieved 99.98% accuracy for
the lower samples of Dental Radio Graphy database and mini-MIAS database. These
architectures are trained for 250 epochs except for MLP which is trained for just 85
epochs to achieve a saturation stage of 99.99% accuracy.
The architectures in the paper will be explained later in the Sect. 2 & Keras
Backgrounds in Sect. 3 and the architectures of our models in Sect. 4, results along
with tensor graphs will be explained in the Sect. 5 and the conclusion, future work (i.e.,
for real time scenarios) will be explained in Sect. 6 following with References. The
main difference between auto encoders and MLPs is that, the generated MSE is
truncated back to the hidden layers in MLP. The weight sharing mechanism between
encoders and decoders in the network enables fast self-learning process but at the same
time it restricts the degree of freedom of the network.

2 DAE, SDAE and MLP

Auto Encoders are self-supervised, feed forward and non-recurrent neural networks
which has a belief in deep learning to find applications in Image Dimensionality
reduction, data denoising, clustering, machine translation and anomaly detection etc.,

2.1 Denoising Auto Encoders (DAE) and Stacked DAE

The other name of Auto encoders are Diablo Networks or Auto Associators which
undergoes self-learning process using feed forward architecture.
They consist of two cascaded networks [7] - The first network is an encoder which
converts the input signal (x) into encoded signal (y) by using a transformation h(x) i.e.,
y = h(x). And the second network is used for reconstructing the signal. Let’s say it (r).
Then the reconstructed signal (r) can be written as a function of r = f(y) = f (h(x)) and
the difference between the original input and the reconstructed output is the error
(e) obtained. This error is also called as Mean Square Error (MSE) which can be greatly
reduced by self-learning in Auto Encoders Network.
Out of these various auto encoders we use denoising auto encoders for pixel-wise
noise removal from the image and the process of self-learning comes by comparing the
neighbouring pixels. In our project, we used stacked denoising auto encoders to remove
noise where stacked means the input of one encoder is given as the output to the next
encoder in order to take and retrieve the complex features of the image for the max-
imum noise removal and original shape retrieval of the image.
From the architectural design of a general denoising auto encoder shown in Fig. 1,
(X) is the input image having features {x1, x2, x3, …., xn} and (h) is the transfor-
mation function which generates the encoded signal (y = h(x)), (w) and (U) are the
input and output weights to and from the transformation function (h(x)). And (r) is the
reconstructed signal which is a function of the encoded signal (y) i.e., r = f(y) = f (h
(x)), as said earlier.
From the Fig. 1 We can rewrite the equation of the encoded signal (y) in the form
of
6 S. S. Kunapuli et al.

Fig. 1. Auto encoder architecture

y ¼ s ðWx þ bÞ ð1Þ

Where b is the biasing function and “s” could be any non-linear function of h(x).
Now the reconstructed signal function (r) can be rewrite in the form of

r ¼ sðW0 y þ b0 Þ ð2Þ

As said earlier, the weight sharing concept of the auto encoder, that is the weights
of the encoder and decoder are simply a transpose of each other, which we can see that
in the Eqs. 1 and 2. Hence it gives a deterministic approach.
The architecture of the stacked denoising auto encoder (SDAE) is best explained by
Gondara [6] in his work on Medical Image Denoising using Conventional Auto
Encoders.

2.2 Multi-Layer Perception Neural Networks (MLP)

Multi-Layer Perception (MLP) neural network is a feed forward architecture consisting
of one or more number of hidden layers. These hidden layers have a different number
of hidden neurons each carrying the same activation function (Fig. 2).

Fig. 2. Multi-layer perception neural network (MLP)

 Enhanced Medical Image De-noising Using Auto Encoders and MLP 7

Since each neuron of the hidden layer is connected to all the neurons of the
previous hidden layers, it is also called as Fully Connected Layers.
Unlike DAE’s or CAE’s, MLP performs Back Propagations, i.e., feed backing the
error signal to the hidden layers to change the weights of the neurons in the hidden
layers. The activity of the neurons can be expressed by an activation function,

y ¼ RðWx þ bÞ ð3Þ

In our presenting work, we used “Sigmoid” Activation Function for each neuron in
the hidden layers to limit its probabilities from {0, 1} (Fig. 3).

Fig. 3. Sigmoid activation function

3 Keras Environments

While designing a neural network using Keras, we have used two types of back-
grounds, they are, Tensorflow and Theano, so as to ensure which Backend gives the
maximum accuracy at which epoch. In our work, we have tested SDAE on mini-MIAS
database and Dental Radio Graphy database on both Keras environments and found
that Theano has achieved a maximum accuracy of 99.97% at 141th epoch and Ten-
sorflow has achieved a slight more accuracy of 99.98% at 152nd epoch. Whereas, MLP
has achieved 99.99% in retrieving the original image at 87th epoch.

4 Proposed Architectures of SDAE and MLP

In this section, we are going to discuss about the architectures of the proposed Stacked
Denoising Auto Encoders and Multi-Layer Perception Algorithms.
As we can see in the Fig. 4 the stacked Denoising architecture consists of one input
layer of size (64, 64, 1) and two fully connected layers having input dimension 64 with
an activity function of Sigmoid to each neurons in the fully connected layers (Fig. 5).
Figure 6 shows the similar architecture of SDAE but the difference is that the
background environment has been changed to Theano instead of Tensorflow. So as the
input dimension has been changed to “1” instead of 64 and the shape of the input image
has been changed to (1, 64, 64). And the rest of the architecture is similar to that of
SDAE using Tensorflow. It’s like a mirror to the original architecture of Tensorflow
(Fig. 7).
In the MLP architecture, we have added some dropout layers when comparing to
SDAEs, as it supports Back propagation of the resulting errors to the hidden layers
8 S. S. Kunapuli et al.

Fig. 4. Stacked denoising auto encoders (SDAE) architecture using Tensorflow background

Fig. 5. Depicts the SDAE tensorflow architecture which was visualized using tensor board.

Fig. 6. Stacked denoising auto encoders (SDAE) architecture using Theano background

making variation weights of the neurons and also restricting the degree of freedom of
the network architecture. Hence, we used Dropouts to regularize the network archi-
tecture (Figs. 8 and 9).
 Enhanced Medical Image De-noising Using Auto Encoders and MLP 9

Fig. 7. Depicts the SDAE Theano architecture which was visualized using Tensor board.

Fig. 8. Proposed multi-layer perception (MLP) neural network.

5 Results

In this section, the performance of our algorithm is compared with the results of the
previous models with the help of a tensor board and images.
10 S. S. Kunapuli et al.

Fig. 9. Depicts MLP denoising architecture using tensor board

5.1 Results of DAE (Backend: TensorFlow)

Noised Images are Taken at l = 0, r = 0.5, p = 1 and trained the network with a batch
size of 128 with “Adadelta” optimizer and loss “binary_crossentropy”.
The output graph of SDAE using Tensorflow architecture is shown in the Fig. 10
continuing with the Noised, Original and Denoised Images at Figs. 11, 12 below.

Fig. 10. Graph visualization for decreasing loss during recovering till the last pixel of the
medical images
 Enhanced Medical Image De-noising Using Auto Encoders and MLP 11

Fig. 11. Depicts Added Gaussian Noise to the Medical images (As we added same amount of
noise as input to the every architecture, we are representing this Fig. 11 as noise image to all the
architectures)

Fig. 12. Depicts the original and retrieved images using SDAE tensorflow architecture. You can
see there is a slight difference between the images which also negotiable in bio-medical
Instrumentation system but yet it yields a fruit-full results.

Even it got 99.99% of accuracy, there is some slight variation of the Denoised from
the original images. So in a case of diagnosing cancers, it could be somewhat crucial
problem (as we are processing the black box). Hence we are undergoing some pre-
processing steps, which will be extended later to our work so as to achieve 100%
accuracy which will be our future scope (Figs. 13, 14, 15 and 16).

Fig. 13. Depicts the SDAE output using Theano architecture (Here you can see that as model is
same but the architecture is different the fall of error will be different but yields very similar
results to the SDAE tensorflow architecture.)
12 S. S. Kunapuli et al.

Fig. 14. Depicts the original and retrieved images using SDAE Theano architecture.

Fig. 15. Depicts the MLP output during denoising (As we can see the difference between these
graphs as how it falls from the peak value while denoising the noisy images.)

Fig. 16. Depicts the original and denoised Images using MLP architecture.

5.2 Comparison of Accuracies with the Existing Models

Image type SSIM
Noisy 0.63
NL means 0.62
Median filter 0.80
CNN DAE(a) 0.89
CNN DAE(b) 0.90
Proposed SDAE Tensorflow Architecture 0.99
(continued)
Another random document with
no related content on Scribd:
kings had twenty thousand war chariots, and ancient writers say that
there were scattered along the Nile from here to Memphis one
hundred stone stables, each large enough to accommodate two
hundred horses.

Avenues of sphinxes guarded the approach to the ancient Egyptian temple.

Between the paws of each of the ram-headed sphinxes at the great temple at
Karnak, Rameses II placed a statue of one of his predecessors.

The Aswan Dam is a huge granite barrier a mile and a quarter long which now
controls the waters of the Nile after centuries of alternate flood and drought, saves
Egypt from famine, and adds millions of acres to her irrigable lands.
 It brings one close to the days of the Scriptures when he can put
his hand on the very same things that were touched by old Pharaoh;
and can visit the temples in which he worshipped, or sit on the
monuments erected in his honour, and look at the tomb in which his
royal bones were laid away. One feels closer still when he can look
at the royal mummy itself and actually see the hardhearted old
heathen almost as he was when alive, as I did at the museum the
other day.
This Pharaoh, Rameses II, was one of the greatest kings of
ancient Egypt. His temples are scattered throughout the Nile valley
and his statues are the largest ever discovered. One was found in
the Nile delta which measures forty-two feet in height, and there are
others sixty-six feet high at Abu Simbel in Nubia, about as far up the
Nile as Chicago is distant from the mouth of the Hudson. They are
seated on thrones and are hewn from the solid rocks. These figures
stand in front of the temple, also cut out of rock. This building is said
to have been erected by him in honour of his favourite wife,
Nefertari, and there are statues of his children about it. These show
that he was very much of a family man, for inscriptions on the
various monuments mention one hundred and sixty-two of his
children by name.
 CHAPTER XVI
THE NILE IN HARNESS

Within a mile or so of the red granite quarries, out of which

Pompey’s Pillar and the obelisks were taken by the ancient
Egyptians, just below the island of Philæ, with its stone temples built
ages ago to the Goddess Isis, far up the Nile valley, on the edge of
Lower Nubia, I write these notes for my American readers. I am in
the heart of the desert, seven hundred miles south of the
Mediterranean Sea, at the point where the great river drops down
over the first cataract. I have come here to describe the Aswan Dam,
which the British built to harness the Nile and thereby save Egypt
from famine.
We all look upon this as the oldest of rivers, but the Nile god of to-
day has many new aspects. For ages he has been ramping and
charging at his own sweet will, but he is now being harnessed and
will have to work in the traces like an old plough mule. In the past he
has been feeding his daughter Egypt or not, as he pleased. He has
sometimes stuffed her to repletion, and at others has held back his
supplies of water and mud, causing a famine. This was the case
during the seven hungry years of Joseph’s time, and the fat years of
that day were undoubtedly produced by high Niles. Such ups and
downs have occurred in Egypt from time to time since the dawn of
her history, and it is only in comparatively recent years that man has
attempted to control the old river and by a system of dams hold back
the waters and let them out over the farms as needed. To master the
Nile has cost many millions of dollars which have gone into building
the great barrages in the lower river, and more important than all, the
mighty dam away up here at Aswan.
Egypt is almost rainless and the Nile gives both land and people
their food and drink. I have already described some of the wonders
of the stream and what it does for Egypt. It rises in Lake Victoria, in
Central Africa, and drops a distance greater than the altitude of the
highest of the Alleghanies before it flows into the Mediterranean Sea.
In the upper part of its course it is known as the White Nile, and this
should be called the main stream of the river. At Khartum, thirteen
hundred and fifty miles from the Mediterranean, the Blue Nile, which
rises in the Abyssinian Mountains, comes in, while about one
hundred and forty miles farther north the Atbara, or Black Nile, which
is also from Abyssinia, joins the main stream. From the mouth of the
Atbara to the sea there is not a tributary of any kind connected with
the river. It ploughs its way through the desert valley, in which it has
built up Egypt, narrowing and widening, until a few miles below
Cairo, where it divides into two great branches and flows off into the
sea.
The volume of the Nile is enormous. At flood times, a billion tons
of water go by at Aswan every day. The river then rises twenty-five
feet at Cairo, thirty-eight feet at Old Thebes, and almost fifty feet at
the first cataract, where I now am. There is so much water that no
dam could hold it, hence all of these great works had to be made so
that the water can be let in and out and allowed to pass through at
will.
It is at flood time that the Nile valley gets its rich feed of Abyssinian
mud. This is brought down in part by the Blue Nile, but more
abundantly by the Atbara, or Black Nile. It is carried by the
inundation all over Egypt and by means of irrigation conducted to
nearly every farm. After the floods subside the muddy waters grow
clear again. The Blue Nile and the Black Nile become almost dry,
and the white water of the main, or Victoria Nile, is about all that
Egypt has. It is this white water that is stored up by the Aswan Dam,
and it feeds the country in much the same way as our irrigation
canals do, with water only and not with a thick mixture of water and
mud as in the times of the overflow.
For thousands of years these rivers have been pouring down
through this Nile valley; but whenever the rains have been scanty in
the highlands of Abyssinia and in Central Africa the main stream has
not been high enough to reach the whole country. Most of the lands
could be inundated only once a year, and if the Nile was especially
low some could have no water at all. By the present system Egypt
has water all the year round, and enough to make it produce two or
three crops every twelve months.
I have been much interested in the irrigation works of the past.
The whole of the Nile valley above Cairo is cut up into a series of
basins. For six hundred or seven hundred miles north of this point
the valley slopes very gradually and, in order to save the water, dikes
have been made across it and embankments run parallel with the
river, turning the whole country into a series of basin-like terraces,
each containing from five thousand to fifteen thousand acres. These
basins, which are often subdivided, are so connected that the water
flows from one to the other until it finally passes out of the lower
basin back into the Nile. When the floods come, the lowest basins
are filled first and then those higher up, until at last all have become
great ponds and Egypt is one vast inland sea cut up by the
embankments and islands upon which the villages stand.
There are many such systems of basins in Upper Egypt, some
large and some small. There are also basins higher up and closer to
the river which are filled with sakiebs or shadoofs. When I tell you
that the fall of this valley from here to Cairo is only seven inches to
the mile you will see how carefully these basins must be graduated
in order to take advantage of the flow of the river. They have to be so
constructed that the water can be drained off as rapidly as it is let on.
As I have already said, the Abyssinian mud contains a great quantity
of salts, and it is just as bad to have too much of it as too little. If the
land is over-watered the salts dissolve from the soil, the over-soaked
land becomes wormy, and the crops are often sown too late. The red
water, or that containing the silt, is allowed to stand just about forty
days. During this time it drops a great deal of sediment and furnishes
enough moisture for the crops.
But the Aswan Dam has so regulated the river flow that the
Egyptian farmer is far less at the mercy of low Niles or high Niles
than in the past. The dam is one of the wonders of modern Egypt. It
is in full sight of me as I sit here on the left bank of the Nile, with the
desert at my back. It looks like a great stone viaduct crossing the
rocky bed of the river, joining the stony hills which wall the Nile on
both sides, and holding back a portion of its mighty waters. It is a
huge granite barrier a mile and a quarter long. There is now a
roadway guarded by walls on its top, and there is a miniature railway,
the cars of which are pushed by men from one end to the other. The
dam serves as a bridge as well, and donkeys, camels, and men are
allowed to pass over it from bank to bank. I crossed on the car at a
cost of twenty-five cents, my motive power being two Arab boys who
trotted behind.
As I came over, I stopped from time to time to examine the
construction. The dam is made of big blocks of red granite as fine as
that of any tombstone in the United States. They are beautifully cut,
and fitted as closely as the walls of a palace. On the upper side or
south face the wall is perpendicular, forming a straight up-and-down
barrier against the waters of the Nile. I climbed down a ladder on that
side at one place almost to the river, and could see that the blocks
are fitted so closely that the cement does not show. The masonry
seems almost one solid stone throughout, with the exception of
where the great sluices are cut, to allow the river to flow through at
the times of the flood, and as the floods subside to shut back the
waters to form the reservoir for the dry season.
There are one hundred and eighty of these sluice gates in the
dam, each of which has steel doors that can be raised or lowered to
allow the whole river to flow through or to hold back as much or as
little as the engineers will. The dam is thus a great stone wall pierced
by these gates.
The Nile never flows over the top of the dam, but always through
the gates and the canal at one side. When the gates are closed
during the dry season, enough water is held back by this structure of
steel and granite to form a lake over one hundred miles long, and
this is let out as needed to supplement the ordinary flow of the river
and give the crops plenty of water all summer through. There is
water enough in the reservoir to give all the families of the United
States all they could use for four or five months, and enough to
supply Great Britain and Ireland the entire year.
 The weight of this water is stupendous and its force inconceivable.
Nevertheless, during the floods fully as much runs through the dam
every day as the whole supply kept back during the dry season; and
the structure had to be made so that it would retain this huge lake
and at flood time let a lake equal to it pass through.
Talk about the Pyramids! The Aswan Dam is far more wonderful
than they are. The Pyramid of Cheops required one hundred
thousand men and over twenty years in its building. The Aswan Dam
was constructed by about eleven thousand men in four years. The
Pyramid of Cheops was made by forced labour and impoverished
the people. The Aswan Dam cost about twelve million dollars and
the men who worked upon it were better paid than any others who
had ever laboured in the valley of the Nile. Moreover, the dam has
meant prosperity for Egypt. It has added to it more than one million
five hundred thousand acres of tillable land and has increased the
value of its crops by over thirteen million dollars per annum. It has
more than paid for its cost every year. Since it has been built the
yearly tax revenues have gained by two million dollars, and the lands
owned by the government have become worth five million dollars
more.
The dam is also more wonderful than the Pyramids in its
construction. Old Cheops is built on the edge of the desert on a solid
stone platform, and is little more than the piling of one stone upon
another. For the Aswan Dam a trench a hundred feet wide and a
hundred feet deep had to be excavated in the granite rock. This was
bedded with concreted rubble to form the substructure upon which
the masonry was raised. The dam itself contains more than a million
tons of granite and about fifteen thousand tons of steel, and the
calculations of the engineers are so exact that they know just how
much every ounce of stone and steel will hold back.
I have had some talks here with the engineer-in-chief of the dam,
and am surprised at the wonderful intelligence bureau that has been
created in connection with the control of the Nile. Its officials know
the exact weight of the river at every hour of the day. They have
telegraphic reports on what the Nile is doing in Abyssinia, in Central
Africa, and in the Sudan. They have dispatches as to every great
rain, and they know to a ton just how fast Lower Egypt is using the
water, so they can tell how much or how little to let out for the farms.
They even estimate the force of the sun on the water and know how
much it drinks up every day. When the reservoir is full Old Sol takes
a million and a half tons from it every twenty-four hours. They know
what the evaporation is, not only at Aswan, but all along the great
stream and throughout its swamps to its source in Lake Victoria.
 The gift of the Nile is not had without work. Fellaheen too poor to own camels or
bullocks lift the river water from level to level and pour it into the irrigation ditches.
 The fellaheen live in villages and go out to work on the farms. The average mud
hut seldom contains more than one or two rooms and is at the mercy of thick
clouds of dust from the road.

I am also amazed at the strength and delicacy of the machinery of

this remarkable structure. The great sluice gates are each as high as
a two-story house, and so wide that you could drive a hay wagon
through them without touching the walls. They are cut right through
the granite dam and are closed or opened by steel doors, which slide
up and down inside the wall on rollers. Upon the top of the dam there
are machines for moving these gates, so made that a child could
operate them. They are equipped to be operated by electricity, but
they are now worked by hand, and this mighty force, so tremendous
that two billion horses would be required to move it, is now controlled
at will by the muscular power of a single man.
This thought was impressive as I sat below the dam, where the
eight central sluices pressed by the millions of tons of water lying
behind them poured forth their mighty flood. I had climbed down the
steps at the north side of the centre of the dam to make a
photograph of the streams flowing through. They come forth with a
rush like that of Niagara and go foaming over the rocks with a force
that might generate thousands of horsepower. The noise is like
thunder and the torrents fairly shake the earth. Each is about fifteen
feet in height and yellow with mud. There were eight such streams of
golden foam at my right, and farther over I could see the spray from
others all dashing through the dam until they met in a yellow frothing
mass several hundred feet below me and rolled onward down the
rocks to Egypt. They flow out with such a force that they tear up the
rocky bed of the Nile, lifting stones weighing many tons and carrying
them some distance down the river. They have done so much
damage of this nature that a cement foundation has now been made
below the dam itself in order to prevent the gouging out of the bed
which would mean the undermining of the main structure.
But the thirsty land and its teeming millions forever clamour for
more water. Even this great Aswan Dam has not nearly solved the
irrigation problem of Egypt. There are always too many would-be
farmers for the watered area. At the present rate of growth, it is
estimated, the population will have increased by the middle of the
century to twenty millions of people, practically all of them dependent
on agriculture, and so on this one river system. The government has
yet more ambitious schemes for hoarding and meting out its
precious waters.
 At Wady Halfa, about two hundred miles up the river from Aswan,
begins the Sudan, which extends for thousands of miles southward.
In controlling this vast territory, Great Britain has hold also of the
upper reaches of the Nile from the south boundary of Egypt proper
into the Great Lakes of Central Africa where the river has its source.
The irrigation works, new dams, and reservoirs planned or building
on the Upper Nile are intended to increase the arable lands not only
in the Sudan but in Egypt as well. The projects which the British
have for the improvement of the Nile will rank as the most daring of
the engineering plans of the century. To carry them out will cost as
much as the Suez Canal, but they will build up fifteen hundred or two
thousand miles south of the Mediterranean Sea, several other
Egypts twice or thrice as rich as the lower Nile valley, each
supporting its millions of people.
The projects include schemes for the regulation of the Great Lakes
on the highlands of Central Africa, to make them serve as reservoirs
for the Nile. They include, also, plans for the embankment of the
tributaries of the White Nile flowing through the great swamps on the
northern slope of the Congo watershed, and the digging of over two
hundred miles of new channel, whereby the main stream of the
White Nile will be greatly shortened and its bed fitted to carrying the
enormous volume of its waters down to Khartum. Another scheme
contemplates the erection of a dam at Lake Tsana, on the highlands
of Abyssinia, which will make that lake a reservoir for the Blue Nile
and enable it to water the fertile plain which lies between the Blue
and White Niles, ending at Khartum.
The great trouble now is that a large part of the waters of the Nile
go to waste, particularly in the swamps of the Sudd region. These
mighty swamps lie on the northern slope of the Congo watershed
and are fed by the branches of the White Nile known as the Bahr el
Jebel, the Bahr el Ghazal, and the Bahr el Zaraf. They begin where
the River Sobat flows into the Nile and form an irregular triangle, the
base running from that point two hundred miles westward, with the
southern apex at Bor, which is two or three hundred miles farther
south. They lie on the bed of what in prehistoric times was a great
lake, and are composed of masses of reeds, papyrus, and other
swamp grasses, so interlaced that they soak up the water like a
huge sponge. Imagine a sponge as big as the State of Indiana, from
two to six feet in thickness, and so situated that it is always filled by
the waters of the Nile and you will have some idea of this region.
This sponge is near the Equator where the tropical sun beats down
upon it, so that steam is always rising. It sucks up the waters of the
Nile and gives them out into the air. The evaporation in the Sudd and
along the courses of the Nile is so great that an amount equal to half
the capacity of the Aswan reservoir is lost every day. In the summer
fully fifty per cent. of the water supplied by the Great Lakes never
gets into the main stream of the Nile. The water of this swamp is
nowhere much above a man’s head, and in most places, except
where the main stream flows through, it is only waist-deep. The
evaporation increases at the time of the flood, when more land is
covered, so that no matter how much water flows into the swamp,
only about the same amount flows out.
The vast masses of floating weeds break up and burst into the
channels, and when an obstruction is encountered they pile up on
one another just as ice does. In the hot, dry season, when the stems
of the papyrus are ten or fifteen feet high, the natives start fires
which sweep the region from end to end, destroying all other
vegetation. The ashes and burnt stems add to the floating mass,
which after a time becomes five or six feet in thickness and almost
like peat.
In clearing this Sudd and reopening the channels, the first step is
to cut down the vegetation. The sponge-like mass is then cut with
long saws into blocks, much as ice is harvested on our ponds. The
blocks are pulled out into the current by steel cables attached to the
engines on the steamers and float down the stream. An immense
deal of this kind of work is going on all along the Upper Nile, for it is
only in this way that navigation is kept open.
I have met some of the surveyors who are breaking a way through
the Sudd. They describe it as a vast sheet of brilliant green made up
of papyrus, feathery reeds, and sword grass. These rise from five to
fifteen feet above the water and are broken here and there by
patches of ambatch trees and by channels, pools, and lagoons. The
greater part of the region has no human inhabitants, especially that
along the Bahr el Ghazal.
Big game is to be seen only to the south of the swamp area. There
the land is a little higher, and elephants, giraffes, and buffaloes
inhabit the edges of the swamps. In the heart of it, in fact, in all parts
of it, there are vast numbers of hippopotami, and there are all sorts
of swamp birds everywhere. From the reeds and the mud banks
clouds of wild cranes, geese, storks, herons, pelicans, and ducks of
every description rise up as the boats approach, and there are
insects by millions—mosquitoes, moths, spiders, and flies. There are
other insects that carry fevers, and the tsetse fly, which causes the
sleeping sickness.
When all the Upper Nile plans and projects have been put through,
the whole river will indeed be a magically powerful, yet tamed and
harnessed, domestic animal at the command of the farmers of a
greater Egypt and a greater Sudan.
 CHAPTER XVII
STEAMING THROUGH THE LAND OF CUSH

For the last two days I have been steaming through one of the
oldest lands of the globe. I have been travelling up the Nile through
the country which belonged to Noah’s grandson, Cush, who was
Ham’s eldest son, and which was known to the Greeks and Romans
in later days as Ethiopia. The Egyptians called it Nubia, from their
word noub, which means gold, and it is known that a large part of the
gold of ancient time came from it.
Ancient Nubia had a considerable population, and was noted for
its riches and power. It was already a flourishing country about the
time of the Pyramid builders, while in the most prosperous days of
Old Egypt it had large towns and magnificent temples dedicated to
the worship of the Egyptian gods. On my way here I passed Abu
Simbel, a great temple on the bank of the Nile, which was cut out of
the rocks by Rameses II, the Pharaoh of the Bible. Farther down the
river lies the Temple of the Lions, where that same old king was
himself worshipped as a god.
Until 1100 b.c. this country was a dependency of the Pharaohs. It
then became independent, and later its armies overran and
conquered Egypt. As other nations came into this part of the Nile
valley they sent their armies against the Nubians, but were driven
back, and at the time the Romans came the country was ruled by a
succession of queens named Candace, one of whom made war
upon the Romans. The Nubian people very early adopted
Christianity, but later, when the Mohammedans took possession of
Egypt and the Upper Nile valley, they were converted to Islam. They
are still followers of the Prophet, and were among the boldest
soldiers of the fanatical Mahdi in his fights against the troops of
Egypt and Great Britain.
 A land with such a history ought to be a rich one. The Nubia of to-
day is about as barren as any country on earth. With the exception of
a narrow band along the Nile, it is altogether desert. Beginning in the
sands of Libya, it extends several hundred miles eastward to the
Red Sea, but only in a few places has the soil enough moisture to
furnish even a scanty pasturage for camels and sheep. The bulk of
the desert population is made up of Bisharin Bedouins, living in tents
made of matting and moving about from place to place with their
flocks. Each tribe has a certain number of wells, and water is the
principal part of its visible wealth. The British officials of the Sudan
have surveyed these wells and investigated their depth and the
quality of the flow of the water. The government has also sunk some
new wells and found water at a depth of about one hundred feet.
Nubia is now a part of the Upper Nile valley, a cultivated strip, in
places only a quarter of a mile wide, winding its way like a snake
from north to south as far as from New York City to Detroit, and
extending on both sides of the river. It is of irregular width, for in
some places the desert comes close to the river, while in others the
stream winds through black rocky hills which rise straight above it a
thousand feet. Farther on, one sees yellow sand, spotted with black
rocks, which show signs of volcanic origin, and then at a low bend in
the river the water may be conducted out over the sands and create
a cultivated patch three miles in width.
The Nile is so walled in by hills that its waters have to be lifted in
order to flow over any level place. This is done chiefly by the
sakiehs, of which there are something like four thousand on the
Nubian Nile. The great wheels, moving in cogs, can be seen high up
on the banks, with their strings of buckets hanging to them. As the
buckets descend, each dips into the water and carries to the top a
few quarts at a time. In some places men raise the water in baskets
or buckets, and in others, the river slopes at such an angle that they
carry it up by hand and water little patches twenty or thirty feet wide.
Every low place along the river is farmed, and when the Nile falls,
the sand banks and islands are planted to crops.
Wherever there is a stretch of cultivated land, a village of mud and
stone huts has grown up, and such villages spot the banks for
hundreds of miles. At times there is no green except between village
and river, and one wonders how men can be born and live and die
there. Nevertheless, there are more than one hundred thousand
people to whom this region is the centre of the world.
Though much of this Nile border is too narrow for profitable
cultivation, it is very fertile and raises excellent cotton. At present the
other chief crops are wheat, barley, and millet, and the chief fruit is
dates, which are sweeter and larger than those grown farther down
the Nile valley. Indeed, the date trees that one sees almost
everywhere along the banks are a source of revenue for the
government, which taxes them at the rate of ten cents per tree.
 “On the Ibis we make about six miles an hour as our dusky Nubian pilot
corkscrews up the Nile. Fortunately we are almost free from the myriad flies, the
modern plague of Egypt.”
 Though the Aswan Dam has been of inestimable benefit to Egypt, the whole
world shares regret that when the sluice gates are closed the water backs up and
submerges Pharaoh’s Bed and other ancient ruins on the Island of Philæ.

The steamer Ibis, on which I have been travelling, is one of the

little vessels of the Sudan government which go twice a week from
Shellal, just above the Aswan Dam, to Wady Halfa, where the