Engineering partnerships

In partnership with Australia-based LBT Innovations (LBT), we created the world’s first AI microbiology screening technology for use in pathology laboratories. 

The Automated Plate Assessment System (APAS) aims to speed up the rate of patient diagnosis and treatment by automating the functions involved in screening culture plates after incubation. APAS takes high-quality images of the plates, analyses and interprets any microbial growth, matches this against key patient data, presents a diagnosis, and continually updates its own knowledge base.

APAS will provide doctors with quicker access to more information, and one day may even allow country hospitals to run their own tests without having to ship samples to a central lab, potentially saving a significant amount of time, and importantly, lives.

Our researchers have developed an autonomous medical image program to detect breast tumours. 

The system is used in conjunction with an MRI scan, and using artificial intelligence employs the traversal movement and style of a retro video game to examine the breast area. 

Just as vintage game Tetris manipulated geometric shapes to fit a space, this program uses a green square to navigate and search over the breast image to locate lesions. The square changes to red in colour if a lesion is detected. 

This unique approach is 1.78 times faster in finding a lesion than existing methods and just as accurate. 

The program was developed using deep reinforcement learning methods, a form of AI that enables computers and machines to learn how to do complex tasks without being programed by humans. As a result the program can independently analyse breast tissue. 

Our ultimate aim is for this detection method to be used by radiologists to complement, support and assist their important work in making a precise and quick prognosis.

Several years of research carried out on large-scale video surveillance led to the formation of Adelaide based company, Snap Network Surveillance. 

The core technology used by Snap focuses on synchronising surveillance camera networks as a whole, rather than individual cameras in isolation. As a result, the network can automatically learn the relationships with cameras across the network and then apply this knowledge to generate a straightforward and intuitive surveillance tool for security operators.  

The technology is able to scale to networks that have thousands of cameras, and can be integrated with sensors.

Quick and efficient measurement to assess vegetation in natural ecosystems around Australia is being made possible by an innovative research project combining computer vision and environmental science.

University of Adelaide and TERN researchers developed a method to automatically determine the biomass (total amount of plant material) by analysis of a set of panoramic photographs taken at wooded sites.

Images were taken at 300 sites around the country in a variety of ecosystems to establish a baseline measure of the condition of Australia's natural vegetation. Computer vision and machine learning techniques were employed to estimate the cross-section of each tree photographed at a standard height of 1.3 metres (known as the basal area measurement). 

The system analyses a set of three 360 degree panoramas from the centre of a site and then processes the images to produce a 3D reconstruction. From that, the system automatically detects individual trees, estimates the diameter at 1.3 metre up the tree trunk and puts them all together for an overall estimate of the biomass of the site.

This is a much more efficient method of biomass estimation than the standard method of direct measurement which is very labour intensive, using laser readings which is expensive, or other rapid measures which only provide crude estimates. The field method is quick accurate, and requires minimal training.

Adelaide and US-based Optimatics is a provider of infrastructure planning software that enables water and wastewater utilities to work efficiently.

Aware of the benefits that machine learning could bring to the organisation, Optimatics approached AIML’s Director of Advanced Reasoning and Learning, Associate Professor Javen Qinfeng Shi, to develop processes to complement their software.

The aim of this preliminary project was to improve the proficiency of water pipeline pressure sensors that are scattered along the pipes to predict leaks. This would ensure that any leaks would be located, and fixed, in a timely and cost effective manner. AIML’s team was successful in creating a machine learning algorithm to detect water leaks, resulting in near 90% accuracy in a very short amount of time.   

Impressed by the quality of the work, Optimatics engaged AIML to develop a deep neural network to bypass existing slow and expensive evaluation system and compare two pipeline designs as which one ‘dominates’ (a technical term in pipeline design). The team from AIML achieved productive results in very short amount of time and provided advice for the next step.

Optimatics and AIML are currently working together on how to use AI to predict superior pipeline designs.