Healthy Society

Effective communication is crucial, especially for neurodivergent individuals who face challenges in expressing themselves. Our project, "Talk For Me," is an Augmentative and Alternative Communication (AAC) app designed to help these individuals communicate more easily. The app leverages Artificial Intelligence (AI), specifically Large Language Models (LLMs), to process images and generate contextual text, offering a more personalized and intuitive communication experience. We've focused on making the app user-friendly, ensuring it's simple and intuitive for anyone to use. The app provides the right words at the right time and adapts to the user's location and situation, making communication more natural and context-aware. Through rigorous testing and research, we've made significant strides in creating a solution that enhances the quality of life for users, enabling smoother and more meaningful interactions. This project not only benefits individual users but also has the potential for broader integration into the healthcare market.

Project by:

• Keefe Zebastian Dela Cruz 
• Addy Dhingra 
• Matthew Fowler  

This project aims to develop new bioreactors to culture skin tissue, thereby advancing the field of burn treatment. The main function of the bioreactor is to automate the efficient and reliable production of tissue skins, which includes improving liquid distribution accuracy, minimizing media residue, and ensuring consistent media mixing. Among them, the entire process is sterile, while integrating sensors for monitoring skin health and automated skin inoculation are also important components of the product. The project acknowledges the likelihood of collaboration among various researchers and underscores the importance of structured prioritization in collaboration with the supervisory committee. The ultimate goal is to markedly enhance burn treatment techniques through the delivery of a dependable and automated solution in skin tissue engineering.

Project by:

• Kaicheng Lan 

Oxidative stress is an established causative factor for the onset of various diseases. While reference standards, e.g. isoprostanes, are used to detect lipid breakdown and considered the gold standard to measure oxidative stress, the methodology of detection is labour intensive and cannot produce rapid results needed for immediate diagnostic value. A point of care test is currently lacking to facilitate medical decisions by providing clinicians with time critical information relevant to implement treatment interventions and protocols.

A novel in vivo oxidative stress (IVOS) biosensor has been developed, providing real-time measurements of carbonylated macromolecules, a different indicator of oxidative stress. This biosensor was validated using isoprostanes as reference standards and can establish oxidative profiles for biological samples ex vivo, living cells (in vitro) and within individual animals or humans (in vivo).

Project by:

• Serena Barnes 

Low alcohol beer (<1.2% ABV) has gained popularity in the drinking market due to higher quality products and more flavourful offerings being available. Various methods of producing low alcohol beer are used in the industry including  mash temperatures at the extreme ends of the usual mash schedule. However, which of these strategies leads to the best quality final product is not certain. The aim of this project was to create a "good quality" low alcohol beer utilising the extreme ends of the mashing temperature spectrum, in which a sensory analysis was conducted to determine the quality of the final product. A specialty IPA utilising the Windsor Yeast was brewed in which batches with a mash temperature of 50oC (low temperature mashing), 78oC (high temperature mashing), and 82oC (high temperature mashing) were made.  

Project by:

• Panos Constantinou 
• Jacob Karnas 
• Damien Cappelluti 

Magnetic Resonance Imaging (MRI) is a radiation-free medical imaging modality, and advanced sequences can now provide "snapshot" images of internal anatomy while a participant performs slow motions in the scanner. This provides an opportunity to study neck motion. However, performing repetitive head flexion movements in a lying position quickly fatigues the neck muscles. The aim of this project was to design an MRI-compatible apparatus to passively support the mass of the head during head-neck motion in the MRI, and to develop an image processing pipeline to measure vertebral motion.

Project by:

• Julia Clark

Cricket batters often suffer hand injuries from the impact of fast-moving balls, which can lead to long-term pain and osteoarthritis. The batting gloves are their primary protection against such injuries, with different styles offering varying levels of defence when the hand is open versus clenched into a fist. This project seeks to compare the peak force of three different glove styles from the same brand and to examine how gloves of the same style from different brands, using diverse materials, perform. A specialised piston device with a cricket ball-sized impactor was used to simulate impacts on gloves fitted to three different collar circumferences. Force sensors under the gloves captured the data, which was then analysed using an adapted software system. The findings will offer insights into the safety performance of these gloves.

Project by:

• Demeter Karagiannis 
• Lili Rothall 
• Sophie McGrath 
• Jordyn Fleetwood 

Cystic fibrosis (CF) is a genetic life-shortening respiratory disease which has no universal cure. The Robinson Research Institute is seeking minimally invasive ways to observe the efficacy of their pre-clinical treatments in rodents. Hence, we have developed a novel translation mechanism to support a miniaturised endoscopic probe for imaging rodent airways. Ex vivo optical coherence tomography (OCT) scans of a rodent trachea showed the probe could visualise ciliary movement under manual handling. A linear translation system was developed using integrated software and hardware for controlled pullback of the probe. A rotary translation mechanism was then incorporated to allow for 3D, luminal scans. Designs underwent safety and performance testing and were assessed with ex vivo rat airways. The images produced from the system indicate that miniaturised OCT probes, with a controlled translation system, has promise for in vivo visualisation of airway cilia to support CF research in rodents.

Project by:

• Mila Calitz 
• Rachel Tang 
• Albert Zanardo

The Cold IPA is a beer style that emerged in 2018, considered to be a very hop-centric style with minimal malty flavours. One of the final stages of the brewing process is fermentation, in which yeast converts most of the sugars extracted from the malt to ethanol. The fermentation temperature can have a drastic impact on the final flavour, with yeast working slower but producing a cleaner taste with fewer off-flavours at lower temperatures. Additionally, some beer styles add hops, which give beer much of its bitterness and fruity flavours, after the fermentation in a process known as dry-hopping. The effect of fermentation and dry-hopping temperatures in the Cold IPA style have not been investigated, with recipes quoting highly variable conditions. This project aims to determine what temperatures result in the best tasting Cold IPA based on blind tasting results.

Project by:

• Samantha Black 
• Aidan Holdsworth 
• Angelina Isler 

Can Ale Yeasts be used to brew Cold IPA's?

The brewing industry is constantly evolving through various experimentations of ingredients to accommodate to the different preferences that beer drinkers have. Recent beer development has been rapid and the industry has seen many innovative creations with unique flavour profiles being introduced to the market. One example of this is the Cold IPA developed by Kevin Davey of Wayfinder Brewery located in the United States. This specific beer style utilises a lager yeast and is fermented at lower temperatures resulting in a clean and crisp finish. Lager yeasts are known to be unstable during fermentation in comparison to ale yeasts that are typically used in beer production, hence this project aimed to discover the effect ale yeast would have on a Cold IPA beer. Three batches of Cold IPA were brewed using the same recipe exlcuding the yeast type, in which a Lager, American Ale and German Ale were employed to investigate any statistically significant results in the final product through performing a sensory analysis.  

Project by:

• Anisha Catts 
• Grace Woods 

In 2021, the World Health Organisation reported that heart attack was the leading cause of death worldwide, with over 100,000 Australians affected annually. A heart attack occurs when plaque builds up inside an artery, blocking blood flow.

This project focuses on heart attack prevention by using a computer-generated model that mimics a real artery. Three different computer-generated models were evaluated throughout this project and the most suitable was identified. The model helps determine where plaque build-up may occur and assess a patient's risk of a heart attack.

The model will enable early detection of plaque build-up, allowing medical professionals to treat heart problems before they become catastrophic; thereby reducing the overall impact of heart attacks on the health system and the wider Australian population.

Project by:

• Emerson Nugent 

By 2050, the World Health Organization estimates that antimicrobial resistance (AMR) could cause 10 million deaths annually, with an economic impact of 100 trillion dollars. AMR is one of the top ten global health threats, with resistant bacteria able to transfer between animals and humans. Given that nearly 50% of households worldwide own a dog, the spread of AMR in dogs is particularly concerning, leaving limited treatment options. In my project, we are partnering with Neoculi to develop new antibiotics that are more effective and have fewer side effects to combat resistant bacteria in dogs. Our aim is to tackle the urgent challenge of antibiotic resistance and provide innovative treatment options for canine infections. This collaboration also seeks to ensure the rapid development and commercialization of these drugs, making them widely accessible.

Project by:

• Sayara Bista

A common surgical procedure to treat femoral shaft fractures uses intramedullary nails which are inserted into the middle of the bone to stabilise the fracture site. The current procedure for inserting these nails exposes both patients and surgical teams to substantial ionising radiation, which is known to increase cancer risk. This project aims to further innovate the Austofix Ezy-Aim device, a radiation-free targeting system for inserting locking screws into the implanted nail. A combination of simulation and mechanical testing has been used to inform the significant design changes made to the Ezy-Aim to increase the accuracy and usability of the system. The results of this project will contribute to ensuring the Ezy-Aim is a safe, reliable option for femoral fracture fixation.

Project by:

• Joshua Sharp 
• Joshua Davis 

Ever wondered how much wine is lost before it even reaches your glass? This project explores hidden inefficiencies within winemaking, aiming to save every precious drop. By investigating where wine losses occur in the process, this research seeks to enhance the efficiency and sustainability of winery operations. Using a combination of data analysis and fundamental chemical engineering principles, key points in the production process where losses are most prevalent are identified. The goal is to develop strategies to minimize these losses, ensuring that more of the product ends up in bottles, not down the drain. This study not only contributes to reducing waste and improving waste quality in winemaking but also supports more sustainable production practices, benefiting both producers and consumers. Findings can pave the way for more resilient and efficient wine production, ultimately enhancing the overall quality and availability of wine.

Project by:

• Alicia Stefanoff

Streptococcus pneumoniae (the pneumococcus) is a harmful bacteria that can cause serious illnesses like ear infections (otitis media), pneumonia, and meningitis. We have vaccines to protect against some types of these bacteria, but other types not covered by the vaccine have started to cause infections. This makes it harder to stop the spread of these diseases.

To understand this problem better, we did a systematic review and looked at studies from different sources. We found that while the vaccines have reduced some types of the pneumococcus, other types not covered by the vaccines are increasing, and some of these are resistant to many antibiotics.

To fight these infections, we need new antibiotics, better use of existing ones, vaccines that work against all types of the pneumococcus, and strong infection control methods.

Project by:

• Gabriel Sunmonu 

Schizophrenia is a serious mental illness. Our project aims to use artificial intelligence (AI) to assist doctors in diagnosing schizophrenia. Imagine that a computer could make a diagnosis of schizophrenia by recording a conversation between a doctor and a patient through a microphone or reading an audio file directly. Our project divided it into three parts: speech to text and extraction of features, extraction of acoustic features from audio, and finally classification of these features into a pre-trained support vector machine (SVM) to get diagnosis results. Doctors can use the system's diagnostic results as a reference and an aid to reach a faster diagnosis. We hope this system will help doctors make faster and more accurate diagnoses and help patients get the treatment they need faster.

Project by:

• Wenfei Zhu 
• Nuoyan Ruan

Natural products have broad applications in the field of pharmaceuticals, contributing to the development of new drugs and bioactive compounds. However, the large-scale production of certain products, such as steroids, is limited by the requirement for multi-step reactions and expensive reagents. To overcome this, the catalytic potential of a class of enzymes found in microorganisms, known as cytochrome P450s, has been explored to simplify these chemical transformations. These enzymes can catalyse efficient chemical reactions in a single step, and can be modified to do so using only hydrogen peroxide as a co-substrate. This project aims to evaluate one such modified P450 peroxygenase enzyme from a Nocardia bacterium. This will be achieved by assessing the enzyme's steroid hydroxylating substrate range and efficiency, as well as its stability to temperature, organic solvent and hydrogen peroxide. These insights will determine the enzyme's capacity for catalysis, and the optimal conditions for larger scale applications.

Project by:

• Annalise Abbott

Stroke causes damage to the brain often resulting in loss of motor function in limbs. More than 100 Australians have a stroke each day, 35% of survivors are left disabled and lose independence in daily living. A major contributor to this is partial paralysis of the hand. Stroke rehabilitation aims to improve limb mobility and control. Active rehabilitation is a key component and involves patients voluntarily activating their muscles to perform exercises. This can be achieved with an assistive exoskeleton that can react to the patient's intention. Many exoskeleton hand designs have been developed, however, these are often expensive and have limited dexterity in the fingers and thumb. This project aims to develop an affordable exoskeleton hand with improved dexterity, particularly for the thumb, which is significant for a range of tasks in daily life. The exoskeleton will be realised using a 3D printed rigid link system with linear actuators.

Project by:

• Minh Duc Do 
• Ashford Tran 
• Aidas Pocius 

After nasal surgery, 10% of patients feel a sensation of medicine entering their ears during nasal rinsing. To understand this phenomenon, we have modelled a patients nasal cavity from CT-scans. Using a computational fluid dynamics software "ANSYS Fluent" we simulate the flow of the medicine throughout the nasal cavity. By doing this, we can understand why the medicine may enter the ear, also, a computer aided design software "ANSYS Spaceclaim" was used to perform virtual surgeries on the nasal cavity model. As a result, we can help doctors advise patients on the best techniques for nasal rinsing to reduce the chance of the medicine reaching the ear. With this knowledge, surgeons can understand how to optimise the surgeries to prevent issues with nasal rinsing.

Project by:

• Daniel Kim 
• Danielle Zafiropoulos 
• Syed Ahmed Luqman 

"AI to Help Doctors Understand X-rays Faster"

In hospitals, doctors often rely on X-rays to figure out what's wrong with a patient. However, looking at X-rays and understanding what they mean takes time, especially when doctors need to carefully describe the results in reports. My project aims to help doctors by creating a smart computer program that can look at X-rays and automatically explain what it finds. The program uses a special type of AI that can learn to recognize different health issues, like heart or lung problems, based on thousands of X-rays and their reports. By training the program with this data, it learns to "see" and "describe" what might be wrong in the patient's X-ray. This could save doctors time, help them focus on treating patients, and make sure everyone gets faster, more accurate diagnoses.

Project by:

• Joel Parakal 

Breast cancer affects over 50,000 Australian women annually and catching it early is crucial for saving lives. Our project focuses on using artificial intelligence (AI) to assist healthcare professionals in detecting and understanding a condition called lactational mastitis, which is linked to breast cancer. The goal is to create an AI model that can analyze protein data from patients and identify patterns that could signal early stages of the disease. By training the AI with data from known proteins, we aim to make diagnosis quicker and more accurate. The result is a smarter, more efficient way to detect the condition early, potentially saving lives and reducing the impact of breast cancer.

Project by:

• Tanjir Ahmed Emon 

Cystic fibrosis (CF) is a serious genetic disease that causes thick, sticky mucus to build up in different parts of the body, like the lungs and digestive system. Doctors currently use tests like X-rays and breathing checks to monitor the disease, but these methods don't give a full picture of what's happening inside the lungs. Our project uses a new imaging technology, X-ray Velocimetry (XV), to capture lung motion as they breathe. We hope to find patterns that show the differences between healthy and CF-affected lungs by applying machine learning techniques. This could provide doctors with better tools to monitor CF and help develop more effective treatments.

Project by:

•  Yi Wang

Our project aims to help doctors predict a rare type of blood cancer called Chronic Myelomonocytic Leukemia (CMML) more accurately. This cancer can be hard to detect early, and our goal is to make it easier and quicker to find, so people can start treatment sooner. We use special computer programs that look at pictures of blood cells and information about patients' health to make predictions. This method is like how a detective gathers clues to solve a mystery.

We collected lots of data, including images of blood cells and other important health details from patients. Then, our computer program learned to notice patterns in this data that might suggest someone has this cancer. Our findings showed that this new way could really improve how doctors figure out who has this disease, making it faster and more reliable. This could help patients start their treatment earlier and have a better chance of getting better.

Project by:

• Zhenzhuo Ye 

In this project, we employ Artificial Intelligence in diagnosing diseases in children by analyzing Medical Images and the reports that accompany the images. In reality, doctors have to visually inspect such images for symptoms of a disease, which requires a lot of time. The goal is to speed up this process by training the computer to recognize the images and the reports. Such reports mention various parts of the body and it becomes challenging to associate them with the correct images. To do this, we have developed a mechanism that allows a computer to read the reports and associate them with the images, while ensuring that it is aware of which body part is being referred to. Ultimately, this will assist doctors in save time and make more accurate diagnoses by automatically detecting diseases from both the image and the text.

Project by:

• Akshay Gole 

Fractures are one of the most widespread health issues globally, with 178 million new fractures and 455 million prevalent cases of acute or long-term symptoms reported in 2019 (Source: Global Burden of Disease Study, 2019). The high number of patients and limited medical resources often lead to longer waiting times for fracture diagnosis and treatment. Our project aims to address this challenge by improving fracture detection using advanced computer technology to analyze 3D computed tomography scans. We employed theoretical models called DenseNet and U-Net to automatically identify and assess fractures from detailed bone images. The result is a system that enhances both the speed and accuracy of fracture detection, helping doctors provide quicker and more precise care.

Project by:

• Xuanqiao Zhang  

Cancer treatments can sometimes cause side effects that make patients feel worse. To help doctors predict these side effects, we are creating a computer program using artificial intelligence (AI). Our project aims to understand the different signs in the body, like genetic information and patient details, to predict who might experience these side effects.

We collected information from patients, including samples like blood and saliva, and looked at their genetic data. By feeding this information into our AI model, it learns to spot patterns that might indicate a higher risk of side effects.

The goal is to help doctors tailor treatments better for each patient, reducing unpleasant side effects and improving their overall well-being.

Project by:

• Cheng Lu 

Plant-based anticancer pharmaceuticals are a safer, eco-friendly and less toxic alternative to conventional treatments. This project focuses on optimising and scaling up the purification and extraction processes of ‘parthenolide’, a key ingredient in an oral anticancer agent derived from a plant called Feverfew. The aims include identifying impurities at various solvent concentrations and utilising separation methods to enhance product purity. By overloading the chromatographic column, we aim to determine the maximum yield and retention of parthenolide. Expected outcomes are the development of protocols for impurity identification, a comprehensive understanding of process parameters, and the creation of a final optimised process ready for industrial upscaling, ultimately improving efficiency and product quality.

Project by:

• Daniel Choo