Projects UoA

Big data collection and sentimental analysis of mobility

during disasters: a case of New Zealand
Supervisor
Minh Kieu
Seosam Costello

Discipline
Civil and Environmental Engineering

Project code: ENG028

This project aims to collect Tweets in New Zealand to analyse individual
evacuation behaviour during and after natural disasters, such as earthquakes,
volcanic eruptions and flooding. New Zealand is one of the riskiest countries in
the world in terms of natural disasters. This project will enhance our
understanding of how people react to disasters and improve our preparedness
for the next catastrophic event.

Analyse human travel patterns during natural disasters

Supervisor
Minh Kieu
Prakash Ranjitka

Discipline
Civil and Environmental Engineering

Project code: ENG036

During and after natural disasters, we often see cities where everyone seems to
be evacuated in a completely inefficient way. This is because human behaviours
during natural disasters are largely still less understood. This project will look at
a Big social media data: 50 million Twitter messages posted before, during, and
after Hurricane Sandy to analyse travel behaviour during natural disasters.

Impaired Automatic Speech Recognition

Supervisor
Reza Shahamiri

Discipline
Electrical, Computer and Software Engineering
Project code: ENG050
Automatic speech recognition (ASR) can be very helpful for speakers who
suffer from dysarthria, a neurological disability that damages the control of
motor speech articulators. ASR can act as a medium to not only understand the
impaired speech but also to talk on thier behaf that enables them to have a
better social and digitial life. Nevertheless, normal speech recognition systems
have not been able to understand dysarthric speech leaving speech impaired
individuals not being able to utilize ASR technolgoies; technologies that can be
life changing for them.
We are looking for candiadtes with background in software enginereing, data
science, computer science, computer engineering, or information technology,
excellent programing skills, good understanding of Machine learning and/or
deep learning, basic understaing of how ASR systems work and related
experience is highly appriciated, self-motivated, can-do attitude, excellent
witing skills.

Detection of Autism Spectrum Disorder using AI

Supervisor
Reza Shahamiri

Discipline
Electrical, Computer and Software Engineering

Project code: ENG051

Autistic spectrum disorder (ASD) is a neurodevelopment condition normally
linked with substantial healthcare costs and time consuming assessments
where early detection of ASD traits can help limit the development of the
condition but the mean age of diagnosis in NZ is 6-7 years - usually 2-3 years
after families/carers had expressed concerns. The optimal window for
delivering treatment to children with ASD is at 2-3 years, which necessitates
identification before two. In addition to be time consuming, clinical diagnosis
has accessibility issues and rely on clinical judgment. Screening tools have
been proposed but they are not being used often because of concerns about
accuracy.
We are looking for candiadtes with background in software enginereing, data
science, computer science, computer engineering, or information technology,
excellent programing skills, good understanding of Machine learning and/or
deep learning, self-motivated, can-do attitude, excellent witing skills.

Developing web-based learning applications

Supervisor
Nasser Giacaman
Discipline
Electrical, Computer and Software Engineering

Project code: ENG053

This project will involve designing and building web-based extensions to a
learning resource. The particular domain selected will be determined closer to
the time, but may include helping students learn mathematics or programming
concepts. To be successful in this project, you should be a strong programmer
confident in using HTML, CSS, JavaScript, React and ideally other web-based
technologies also.

Creating a flexible bilingual Text to Speech platform

Supervisor
Catherine Watson, Jesin James, Isabella Shields, Peter Keegan

Discipline
Electrical, Computer and Software Engineering

Project code: ENG054

We have a created a bilingual text to speech system that can speak both te reo
Māori, and New Zealand English – the official languages of Aotearoa.
Developing a text to speech synthesis system involves speech data collection,
data preparation, feature extraction, modelling and using the model to
synthesise speech. Currently, we have implemented this process using the
MARYTTS speech engine which uses the Hidden Markov model. However, there
are other speech engines that use different modelling approaches like deep
neural networks.

The current text to speech system using one speech engine is a client-server
model, enabling anyone on the UoA network to access the system via their
browser. We wish to extend this system to include multiple speech engines.
Such an extended system will help future researchers to test different speech
engines easily and work towards improving the system. This project will involve
creating such a platform. It will require work in both the back end (developing
text to speech synthesiser using different speech engines) and the front end
(combining it all into a usable system). The successful candidate will learn
different methods of doing speech synthesis, acquire speech processing skills,
and gain experience in developing client-server applications.

The ideal candidate will be a confident coder, have an interest in speech, have
knowledge of client/server systems, and human computer interfaces and be
comfortable both working in a group and by themselves. The ability to speak te
reo Māori is not required, but an interest in the language is encouraged.
Telescope scheduling
Supervisor
A/Prof Oliver Sinnen

Discipline
Electrical, Computer and Software Engineering

Project code: ENG055

Telescopes observe celestial objects (or sources) in the sky. The telescopes
are usually movable and can point at different positions in the sky. Very large
scientific telescopes in astronomical observatories are highly specialised, very
expensive and scarce. Hence, they must be used as efficiently as possible.

In a typical operation mode, a telescope will be configured to observe a given

list of sources, which correspond to positions in the sky. It remains pointed at a
source of the list for a certain observation period, e.g. 10 minutes, after which
the telescope is rotated and tilted to point to the next object on the list. Due to
the fact that the movement between positions takes time and that necessary
observation periods might vary, this observation of sources becomes a difficult
optimisation problem.

In this project you will develop scheduling and ordering algorithms to optimise
the use of telescopes. This includes the development of a tool that can process
lists of source to be observed which then creates optimised observation
sequences that can be used for the operation of real telescopes.

Sign-language understanding using Reinforcement

learning and DNN
Supervisor
Ho Seok AHN

Discipline
Electrical, Computer and Software Engineering

Project code: ENG058

This project will understand English sign-language from hand motion, gesture,
and facial emotion. There are a lot of research on recognizing alphabets and
numbers, but not word and sentence itself, which is more realistic and useful in
real-world. Do literature survey on conventional approach, and select the best
one and develop it as the first version. Then improve the method as the second
version and/or integrate them to understand nuance of sign language. You may
use reinforcement learning and/or Deep Neural Network (DNN). This project
may be combined with other related projects, such as social human robot
interaction projects and healthcare robot projects. All necessary devices will be
provided.

This project includes to:

● do literature survey on conventional approach
● find the best one for English,
● develop and test, and
● improve the method
Project scope will be decided after the meeting with supervisors. You will bring
to the role a passion for research and engineering, excellent computing skills
(including a high level of programming ability), and a strong sense of
responsibility. It will be better if you have an experience on vision processing
and/or using DNN.

Plant monitoring system

Supervisor
Ho Seok AHN

Discipline
Electrical, Computer and Software Engineering

Project code: ENG060

This project will develop a system to monitor the change of plant growing. 3D
reconstructed model may be needed to measure the number and size of fruit
and leaves. May need to train Deep Neural Network (DNN) models to detect
different fruits, such as kiwi, apple, grape, strawberry, tomato, paprica. We
have a well working detection model for kiwi and apple fruits, and can easily
make different models for other fruits using DNN. All necessary devices will be
provided.
This project includes to:
● do 3D reconstruction of plant,
● make different models to detect different fruit and leaves,
● find easy way to train models, and
● develop a dashboard for user
Project scope will be decided after the meeting with supervisors. You will bring
to the role a passion for research and engineering, excellent computing skills
(including a high level of programming ability), and a strong sense of
responsibility. It will be better if you have an experience on vision processing
and/or using DNN.
Interactive chatbot system for social robot
Supervisor
Ho Seok AHN

Discipline
Electrical, Computer and Software Engineering

Project code: ENG061

This project will develop an interactive chatbot system for social robot that talk
with visitors at the university reception. We have a working version using
DialogFlow, and will develop a better chatbot. You may use reinforcement
learning and/or Deep Neural Network (DNN) if needed. So when the text is
given, chatbot generates its reactive speech by considering history of
conversation. This project is related to ongoing 5 years research project, SHRI,
and will work with PhD and professional staff. All necessary devices will be
provided.

This project includes to:

● do literature survey on interactive conversation of human,
● find the way to consider relation and history of conversation,
● develop a chatbot, and
● apply it to one of robot platform if possible

Project scope will be decided after the meeting with supervisors. You will bring
to the role a passion for research and engineering, excellent computing skills
(including a high level of programming ability), and a strong sense of
responsibility.

Automated Trustworthiness Testing of Machine

Learning Models
Supervisor
Valerio Terragni

Discipline
Electrical, Computer and Software Engineering

Project code: ENG064

Explainability (or Interpretability) of Machine Learning (ML) models has been
shown to be very important to evaluate the trustworthiness of ML models.
Current approaches automatically identify which aspects of the data influenced
the decision of a ML model (so-called explanations). Evaluating such
explanations usually requires the human in the loop. In this research project, we
want to investigate the possibility of constructing an additional meta-model
that automatically classifies the ML explanations as trustworthy or not. Thus,
enabling automated testing of the trustworthiness of ML models. To be
successful in this project, you should be knowledgeable in supervised ML.
Being knowledgeable in the Explainability/Interpretability of ML is a plus.

Robotic Yield estimation of Apple Orchards

Supervisor
Henry Williams, Trevor Gee

Discipline
Electrical, Computer and Software Engineering

Project code: ENG065

This project will form a part of the multi-institution MaaraTech project being run
by the CARES team at UoA.

The students will process RGBD data collected from Apple Orchard's via a UR5
robotic arm and seek to generate a detailed 3D map of the trees and a count of
the number of fruitless on them with their size, colour, shape, and damage
assesments. This project will involve developing the deep learning-based
computer vision models to detect the fruitless and other canopy features, and
developing methods for 3D reconstruction of the point cloud data.

The research aim will be to extend the computer vision models to handle the
RGBD data, and determine how to associate partially occluded fruitlets across
the scan data to assemble an accurate count for the tree as a whole. This work
will be tested in a real-world field trial with the rest of project team.

This project will primarily be done using Python and/or C++ within a Robotic
Operating System (ROS) framework - utilising pytorch or tensorflow for
development of machine vision algorithms.

===========================
SCIENCE DOMAINS

A behavioural plugin tool for fake news identification

Supervisor
Dr Nalin Asanka Gamagedara Arachchilage

Discipline
Computer Science
Project code: SCI056
Political scientists and researchers are concerned about the impact of fake
news (a.k.a. a combination of misinformation and disinformation) on
democracy. On the other hand, fake news could also escalate to life-
threatening problems, for example, UNESCO states that during COVID-19
pandemic fake news is putting lives at risk, like Coronavirus (COVID-19)
pandemic. Fake news continues to spread, so does people's behaviour and
emotions about the fake news via social media platforms. Fake news via social
media platforms. This opens up the back door for cyber-criminals to entice
people (i.e., taking advantage of victims' emotional and behavioural aspects) to
click on links (e.g., phishing links) associated with fake news when reading.
Therefore, in this project we investigate how people's emotional and
behavioural features influence reading and diffusing fake news in social media
and proposes a fake news detection model incorporating people's behavioural
features and their emotions to better detect fake news in social media.
Furthermore, we study how individuals are emotionally manipulated to perform
various malicious behavioural activities such as clicking on a link associated
with fake news. The project will also explore fake news diffusion patterns and
attacker strategies in social media. The research findings can be used in a risk
prediction model to predict the risk score of any news content being fake in
social media. Eventually, a browser plugin can be implemented through
persuasive design principles for social media platforms that alert the potential
risks to users when dealing with (fake) news.
Prerequisites:
Some experience in programming or Machine Learning (ML)/Natural Language
Processing (NLP) would prefer but not mandatory. Students who have hands-
on experience and skills in Psychology, HCI/game design, rapid prototyping and
evaluation approaches are certainly welcome. Strong interest in human factors
in cyber security/usable security engineering is good, too.

AI and Hyper-spectral imaging for plastic sorting

Supervisor
Patrice Delmas
Martin Urschler

Discipline
Computer Science

Project code: SCI059

New Zealand only sorts about 60% of its plastics while existing recycling
capabilities to do better exist. This project will investigate the potential of AI
and Hyperspectral imaging to separate the seven recognised classes of
recyclable plastics. The SRS student will make use of the IVSLab (https://
ivs.wordpress.fos.auckland.ac.nz/) imaging capabilities to investigate plastics
response to a wide spectrum of light, create an adhoc database and test state-
of-the-art AI algorithms to propose an AI-based plastics sorting strategy.

Characterization and Analysis of Reddit Financial

Communities
Supervisor
Aniket Mahanti
SHOVIK Ghosh

Discipline
Computer Science

Project code: SCI064

The recent surge in prices of meme stocks such as Gamestop and AMC
initiated by regular investor members of the Wallstreetbets Reddit community
has highlighted the power of social media. This project will characterize the
usage of the Wallstreetbets subreddit by using web scraping techniques. Basic
characteristics such as number of posts, comments, and users will be
measured. We will also study a variety of features such as the density of author
posts and timeline of posts and comments over the period of existence of the
subreddit. We will also analyze the stocks that have been discussed by the
community and its impact on the stock price. Data needs to be scraped from
Reddit (https://www.reddit.com/). The data can be collected by tools available
online such as https://towardsdatascience.com/scraping-reddit-
data-1c0af3040768. Stock prices can be downloaded from Yahoo finance.
Skills needed: Python or similar programming, basic statistical analysis, web
scraping, natural language processing.

Understand the usage and popularity of user-

generated video content channels
Supervisor
Aniket Mahanti
SHOVIK Ghosh

Discipline
Computer Science

Project code: SCI065

Youtube is a well-known example of a user-generated video content site.
Similar to YouTube there are many regional and niche video streaming services.
This project aims to characterize such streaming services. We would analyze a
variety of characteristics such as number of videos, comments, likes, views,
and subscribers and various phenomena. The choice of which streaming
service is open for discussion. Some channel data already exists, but more data
can be collected using Web scraping.
Skills needed: Python or similar programming, basic statistical analysis, Web
scraping, visualization.

Network attack prediction using machine learning

Supervisor
Aniket Mahanti

Discipline
Computer Science

Project code: SCI066

We will study the use of Machine Learning algorithms such as SVM, Decision
Tree, and Random Forest in detecting denial of service attacks in networks.
This project will focus on the software-defined network paradigm with a central
viewpoint. The analysis can be performed via simulation such as Mininet or on
existing freely available datasets. The objective is to identify features that can
classify the attacks and also compare the performance of various machine
learning techniques. This project can also be modified to use deep learning
techniques and also the network environment can be changed to IoT networks
instead of the Internet.
Skills needed: Good programming skills, TCP/IP, Mininet or similar simulator,
machine learning.

Recognizing sign language gestures

Supervisor
Mano Manoharan

Discipline
Computer Science

Project code: SCI067

This project investigates the mechanisms for mobile applications to recognize
sign language gestures and convert them to sign language glossing (i.e., the
text representation of sign language) and translate the sign language glossing
to English. This requires investigating some public sign language datasets with
the aim of selecting one or several suitable datasets to be used in the project. It
then requires the use deep learning models to extract the features from the
images and the videos in the datasets and build neural networks generate sign
language glossing, utilising the extracted features.
Python programming skill is essential for the project. During the project, the
student is expected to acquire the knowledge of building neural networks and
using some existing deep learning networks, e.g. GoogLeNet, ResNet, etc., to
extract information from images and videos.