Skip to main content
 Dushani Imesha Perera

Dushani Imesha Perera

Research student,

Email
pereraud@cardiff.ac.uk
Campuses
Room 5.02, Abacws, Senghennydd Road, Cathays, Cardiff, CF24 4AG

Overview

I have a strong passion for research related to Human-Computer Interaction, Human Factors in visualisation, Human Data Interaction, Technology-driven education, Social Science, Virtual Reality, and the Social Dimension of Technology. I believe that a breakthrough in technology can be used in an everyday product providing benefits to society for centuries.

As an undergraduate, I followed Computer Science and Software Engineering at the University of Colombo School of Computing (UCSC), Sri Lanka. I graduated in February 2020 with a Bachelor of Honours degree in Software Engineering from the University of Colombo, Sri Lanka. My final year research project was awarded as the best research for the year 2019. Upon completing my undergraduate studies, I had the privilege of joining UCSC staff and working as an Instructor and then as an Assistant Lecturer. I plan to continue my desire to solve societal problems using ethical & conscientious use of computer, data and information technologies. My hobbies are compering / EmCeeing at conferences and public speaking.

My multidisciplinary research project involves information visualisation surrounding the global climate change issue with the aim of triggering behaviour change.

Research

Research interests

My research interests are in the areas of Human-Computer Interaction, Human Factors in visualisation, Human Data Interaction, Technology-driven education, Social Science, Virtual Reality, and the Social Dimension of Technology. These interests have led me to conduct interdisciplinary projects which include:

  • Drone Pilot Training Simulation Research - Research carried out in developing a drone simulation to train drone pilots in simulated environments. Algorithms using a machine learning approach gave better results in the dynamic model to predict the path over the mathematical dynamic model.
  • An Open-Source Adaptable Extensible Audio Music Streaming and Recommender Platform for Alternate Reality Applications - For my final-year undergraduate research I developed a mobile music platform (Python, Android, and Node.js) to generate auto-completion search suggestions, personalized recommendations, gather user analytics, stream music, and cache songs. I developed a plug-in for Amazon Alexa to play music from the application. Research in the disciplines of Human-Computer Interaction, recommender systems, and auto-completion search suggestions (URL to the codebase: https://github.com/helabeat).
  • Insurance Policy Lapse or Retainment Prediction – I developed a model to detect the lapse or retainment of insurance policies (Keras and Weka).
  • Natural Language Processing: Hate Speech Detection Project - Hate speech detection model (in Python) for Sinhala and English speech in Twitter. More than 60% accuracy was obtained (Logistic regression, n-gram, and word2vec model).
  • Image Processing Application for Sleep Detection of Drivers – I developed a system for sleep detection using Haar-cascades (Python).
  • Fraud Detection in Finance - Several approaches were tried to train and test a model using 4000 unique values for the ‘Synthetic data from a financial payment system’ dataset of Kaggle (the findings include that XGBoost gives better results over KNN and Autoencoders with a test accuracy of 91% obtained on the dataset).
  • ARGAN - Use machine learning and Augmented Reality to deliver designs in the fashion industry.

Thesis

Visualising Household Climate Actions to Trigger Behavioural Change

My multidisciplinary research project involves information visualisation surrounding the global climate change issue.


Climate change is predicted to have significant consequences for the global economy and many sectors of human activity. The objective of limiting the increase in global average temperature to 1.5° Celsius above pre-industrial levels can only be reached if all people, including households, take necessary actions to alleviate climate change. It is evident that families are accountable for a major portion of worldwide Greenhouse Gas (GHG) emissions due to the occupants’ consumption behaviours. Therefore, it is important to study the impact and ways of changing occupants’ behaviours in energy consumption of buildings to reduce energy usage. In view of this, we should explore solutions drawing links between the realms of HCI, behaviour change and climate action for a sustainable environment. We aim to utilise information visualisation to prompt households to take unconflicted climate actions. 


Supervisors

alt

Dr Julie Gwilliam

Dean of Postgraduate Studies College of Physical Sciences & Engineering

alt

Dr Nervo Verdezoto Dias

Senior Lecturer

Publications

2021

2020

  • Perera, D., Rajaratne, M., Sandaruwan, D. and Kodikara, N. 2020. Fraud detection in a financial payment system. Presented at: 3rd International Conference on Human Interaction and Emerging Technologies (IHIET 2020), Paris, France, 27-29 August 2020 Presented at Ahram, T. et al. eds.Multisensory Augmented Reality 2021 Virtual Proceedings, Vol. 1253. Multisensory Augmented Reality 2021 Virtual Proceedings International Conference on Human Interaction and Emerging Technologies: Springer Verlag pp. 520-526., (10.1007/978-3-030-55307-4_79)
  • Perera, D., Raharatne, M. and Arunathilake, S. 2020. HelaBeat: an extensible audio streaming mobile application. Presented at: 3rd International Conference on Human Interaction and Emerging Technologies (IHIET 2020), Paris, France, 27-29 August 2020 Presented at Ahram, T. et al. eds.Human Interaction, Emerging Technologies and Future Applications III, Vol. 1253. Multisensory Augmented Reality 2021 Virtual Proceedings International Conference on Human Interaction and Emerging Technologies: Springer Verlag pp. 115-121., (10.1007/978-3-030-55307-4_18)
  • Perera, D., Rajaratne, M., Arunathilake, S., Karunanayaka, K. and Liyanage, B. 2020. A critical analysis of music recommendation systems and new perspectives. Presented at: 2nd International Conference on Human Interaction and Emerging Technologies (IHIET-AI 2020), Lausanne, Switzerland, 23-25 April 2020 Presented at Ahram, T. et al. eds.Human Interaction, Emerging Technologies and Future Applications II, Vol. 1152. Human Interaction, Emerging Technologies and Future Applications II International Conference on Human Interaction and Emerging Technologies: Springer Verlag pp. 82-87., (10.1007/978-3-030-44267-5_12)