Defense against adversarial attacks on deep learning in computer vision. Computer vision applications rely heavily on deep learning, which is highly vulnerable to being fooled by adding subtle perturbations to object/image textures that are imperceptible to humans. This project aims to develop defense mechanisms to detect and remove adversarial patterns from the input images. The project expects to advance knowledge in understanding the vulnerabilities of deep learning, and to design deep learni ....Defense against adversarial attacks on deep learning in computer vision. Computer vision applications rely heavily on deep learning, which is highly vulnerable to being fooled by adding subtle perturbations to object/image textures that are imperceptible to humans. This project aims to develop defense mechanisms to detect and remove adversarial patterns from the input images. The project expects to advance knowledge in understanding the vulnerabilities of deep learning, and to design deep learning architectures that are inherently robust. The outcomes of this project will increase the security and reliability of computer vision by detecting, reporting and nullifying such attacks and will benefit the general public and industry on many fronts.Read moreRead less
Robust and Explainable 3D Computer Vision. Computer vision is increasingly relying on deep learning which is fragile, opaque and fails catastrophically without warning. This project aims to address these problems by developing new theory in graph representation of 3D geometric and image data, hierarchical graph simplification and novel modules designed specifically for deep learning over geometric graphs. Using these modules, it aims to design graph convolutional network architectures for self-s ....Robust and Explainable 3D Computer Vision. Computer vision is increasingly relying on deep learning which is fragile, opaque and fails catastrophically without warning. This project aims to address these problems by developing new theory in graph representation of 3D geometric and image data, hierarchical graph simplification and novel modules designed specifically for deep learning over geometric graphs. Using these modules, it aims to design graph convolutional network architectures for self-supervised learning that are robust to failures and provide explainable decisions for object detection and scene segmentation. The outcomes are expected to advance theory in robust deep learning and benefit 3D mapping, surveying, infrastructure monitoring, transport and robotics industries.Read moreRead less
Crowd tracking and visual analytics for rapidly deployable imaging devices. Crowd tracking and visual analytics for rapidly deployable imaging devices. This project aims to develop visual analytics technology that adds machine intelligence to a rapidly deployable time-lapse imaging platform. Such devices can operate on solar and wind power, and be remotely programmed (via a cellular network) to take photos and send them to a server at given times. This project, which focuses on monitoring crowds ....Crowd tracking and visual analytics for rapidly deployable imaging devices. Crowd tracking and visual analytics for rapidly deployable imaging devices. This project aims to develop visual analytics technology that adds machine intelligence to a rapidly deployable time-lapse imaging platform. Such devices can operate on solar and wind power, and be remotely programmed (via a cellular network) to take photos and send them to a server at given times. This project, which focuses on monitoring crowds of objects of interest, is expected to introduce “smart” imaging platforms that could be triggered and shoot high-quality photographs when “events of interest” occur. This project could make Australia both a world leader in video analytics and secure through on-line threat detection, and improve traffic control and agriculture.Read moreRead less
I sense, therefore I help: Towards homes that sense and support the aged and infirm. The overall goal is to produce technologies that will enable the home to be a ?intelligent, caring advisor? and provide operational effectiveness for people with decreasing functional capacity (eg aged). It can monitor and support activities of varying complexity without compromising a normal lifestyle, enabling people of varying abilities to be independent, whilst being cared for by their homes. This caring wi ....I sense, therefore I help: Towards homes that sense and support the aged and infirm. The overall goal is to produce technologies that will enable the home to be a ?intelligent, caring advisor? and provide operational effectiveness for people with decreasing functional capacity (eg aged). It can monitor and support activities of varying complexity without compromising a normal lifestyle, enabling people of varying abilities to be independent, whilst being cared for by their homes. This caring will range from advice systems for detecting hazardous situations and alerting the user, through to detecting subtle deviations from complex, normal activities over significant periods of time (such as caused by the onset of an illness).Read moreRead less
Improving the face of cosmetic medicine - an automatic three-dimensional facial analysis system for facial rejuvenation. 'How will I look?' is the most common question to cosmetic doctors from patients considering facial rejuvenation. This project will answer this question for the first time by providing patients with a three-dimensional model of their post-treatment face as well as informing cosmetic doctors exactly how to achieve the patient's desired face.
View and shape invariant modeling of human actions for smart surveillance. This project aims to enable surveillance cameras to interpret videos and detect unexpected activity in real time. Existing surveillance cameras are unable to interpret videos. Because most are not monitored in real time, they play no role in improving security response time. The project plans to develop algorithms to detect actions from any camera viewpoint in continuous videos, a capability that is imperative for smart s ....View and shape invariant modeling of human actions for smart surveillance. This project aims to enable surveillance cameras to interpret videos and detect unexpected activity in real time. Existing surveillance cameras are unable to interpret videos. Because most are not monitored in real time, they play no role in improving security response time. The project plans to develop algorithms to detect actions from any camera viewpoint in continuous videos, a capability that is imperative for smart surveillance yet missing in current techniques. This would improve security and safety response time to events that need immediate attention, such as crimes and medical emergencies, and offer autonomous aids to elderly care, smart homes, child minding, patient monitoring and post-trauma rehabilitation.Read moreRead less
Intelligent Virtual Human Companions. This research aims to develop intelligent virtual human companions that can seemingly integrate our immediate physical environment and understand their surroundings including people’s emotions, behaviours, actions and interactions. Such a technology will be enabled by leveraging recent advances in mixed/augmented reality technologies, and by developing innovative artificial intelligence and computer vision and graphics algorithms for dynamic real-world envir ....Intelligent Virtual Human Companions. This research aims to develop intelligent virtual human companions that can seemingly integrate our immediate physical environment and understand their surroundings including people’s emotions, behaviours, actions and interactions. Such a technology will be enabled by leveraging recent advances in mixed/augmented reality technologies, and by developing innovative artificial intelligence and computer vision and graphics algorithms for dynamic real-world environments. Unlike robots, the proposed technology will be low cost, readily deployable and customisable, and will not have any physical limitations or maintenance requirements. It will thus have a wide range of applications from elderly care, healthcare care to educational training.Read moreRead less
Machine Learning for Fracture Risk Assessment from Simple Radiography. This project aims to develop a novel, reliable, low-cost system to detect poor bone health and assess fracture risk to help to prevent and manage osteoporosis-related fractures. Currently, osteoporosis-related fractures cost our health system millions of dollars annually and costs are increasing with our ageing population. Early detection of poor bone health will improve the effectiveness of preventive measures and ease this ....Machine Learning for Fracture Risk Assessment from Simple Radiography. This project aims to develop a novel, reliable, low-cost system to detect poor bone health and assess fracture risk to help to prevent and manage osteoporosis-related fractures. Currently, osteoporosis-related fractures cost our health system millions of dollars annually and costs are increasing with our ageing population. Early detection of poor bone health will improve the effectiveness of preventive measures and ease this burden. Current methods include unreliable, crude clinical and visual guides that suggest osteoporosis screening. The project plans to develop a novel system by applying machine learning algorithms to radiology data which is commonly captured for diagnosing other conditions.Read moreRead less
Advanced three-dimensional Computer Vision Algorithms for 'Find and Grasp' Future Robots. This project addresses crucial limitations of existing vision systems for the robot grasping of irregular objects in messy living environments. This project aims to undertake fundamental research into novel three-dimensional vision algorithms, exploiting multiple modalities (two-dimensional+three-dimensional+video) for scene labelling, object classification, scene segmentation and grasp synthesis to enable ....Advanced three-dimensional Computer Vision Algorithms for 'Find and Grasp' Future Robots. This project addresses crucial limitations of existing vision systems for the robot grasping of irregular objects in messy living environments. This project aims to undertake fundamental research into novel three-dimensional vision algorithms, exploiting multiple modalities (two-dimensional+three-dimensional+video) for scene labelling, object classification, scene segmentation and grasp synthesis to enable future robots to operate in unstructured environments with highly occluded and cluttered objects. It is expected to significantly advance research and to have broad applications, including home robotics to improve the quality of life of elders and people with special needs. These algorithms may also be used in security (explosive manipulation) and agriculture (field crop harvesting).Read moreRead less
A three dimensional video-based vision system for future robots. With the recent introduction of new three dimensional (3D) video sensors, opportunities for the development of advanced 3D vision systems for robots working in dynamic environments are now becoming possible. A real-time visual robotic system will be developed to substantially reduce the expensive costs associated with elder's health and home care expenses.