ARC Centre of Excellence for Robotic Vision. Robots are vital to Australia's future prosperity in the face of high relative wages, low or decreasing productivity, and impending labour shortages. However the work and workplaces of our most important industries are unstructured and changeable and current robots are challenged by their inability to quickly, safely and reliably "see" and "understand" what is around them. The Centre's research will create the fundamental science and technologies th ....ARC Centre of Excellence for Robotic Vision. Robots are vital to Australia's future prosperity in the face of high relative wages, low or decreasing productivity, and impending labour shortages. However the work and workplaces of our most important industries are unstructured and changeable and current robots are challenged by their inability to quickly, safely and reliably "see" and "understand" what is around them. The Centre's research will create the fundamental science and technologies that will allow robots to see as we do, and overcome the last barrier to the ubiquitous deployment of robots into society for the benefit of all.Read moreRead less
Whole image understanding by convolutions on graphs. This project seeks to develop technologies that will help computer vision interpret the whole visible scene, rather than just some of the objects therein. Existing automated methods for understanding images perform well at recognising specific objects in canonical poses, but the problem of whole image interpretation is far more challenging. Convolutional neural networks (CNN) have underpinned recent progress in object recognition, but whole-im ....Whole image understanding by convolutions on graphs. This project seeks to develop technologies that will help computer vision interpret the whole visible scene, rather than just some of the objects therein. Existing automated methods for understanding images perform well at recognising specific objects in canonical poses, but the problem of whole image interpretation is far more challenging. Convolutional neural networks (CNN) have underpinned recent progress in object recognition, but whole-image understanding cannot be tackled similarly because the number of possible combinations of objects is too large. The project thus proposes a graph-based generalisation of the CNN approach which allows scene structure to be learned explicitly. This would represent an important step towards providing computers with robust vision, allowing them to interact with their environment.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE170101259
Funder
Australian Research Council
Funding Amount
$360,000.00
Summary
Zero-shot and few-shot learning with deep knowledge transfer. This project aims to develop few-shot and zero-shot learning, visual recognition techniques that can learn a visual concept with few or no visual examples. Visual recognition is a major component in Artificial Intelligence and used in cybernetic security, robotic vision and medical image analysis. This project will use deep learning to enable the zero/few-shot learning to use and model previously unexplored information, making zero/fe ....Zero-shot and few-shot learning with deep knowledge transfer. This project aims to develop few-shot and zero-shot learning, visual recognition techniques that can learn a visual concept with few or no visual examples. Visual recognition is a major component in Artificial Intelligence and used in cybernetic security, robotic vision and medical image analysis. This project will use deep learning to enable the zero/few-shot learning to use and model previously unexplored information, making zero/few-shot learning more practical, scalable and flexible. The project is expected to advance the applicability of visual recognition in many challenging scenarios and provide effective tools to analyse the online visual data for supporting Australia’s cybernetic security.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE120101161
Funder
Australian Research Council
Funding Amount
$375,000.00
Summary
Compressive sensing based probabilistic graphical models (PGM). The aim of the project is to develop fast, large scale probabilistic graphical models (PGM) learning and inference methods. The resulting system will be able to process large scale PGMs on a standard PC, and will be easily extendable to computer clustering for larger scale PGMs requiring higher precision.
Statistical Methods of Model Fitting and Segmentation in Computer Vision. Electronic sensors such as cameras and lasers can provide a rich source of information about the position, shape, and motion of objects around us. However, to extract this information in a reliable, automatic, and accurate way requires a sophisticated statistical theory of the process. Example applications include: video surveillance (better automatic detection of moving people and vehicles and of characterising what those ....Statistical Methods of Model Fitting and Segmentation in Computer Vision. Electronic sensors such as cameras and lasers can provide a rich source of information about the position, shape, and motion of objects around us. However, to extract this information in a reliable, automatic, and accurate way requires a sophisticated statistical theory of the process. Example applications include: video surveillance (better automatic detection of moving people and vehicles and of characterising what those people and vehicles are doing), industrial prototyping and inspection (measuring the size and shape of objects), urban planning (laser scanning streetscapes to create computer models of cities), entertainment industry (movie special effects and games), etc. Read moreRead less
Linkage Infrastructure, Equipment And Facilities - Grant ID: LE160100090
Funder
Australian Research Council
Funding Amount
$250,000.00
Summary
Computational infrastructure for developing deep machine learning models. Computational infrastructure for developing deep machine learning models:
The computational infrastructure for developing deep machine learning models aims to enable new developments in machine learning of deep neural network models by providing the specialised computing necessary to train and evaluate the networks. In the last three years, deep networks have smashed previous performance ceilings for tasks such as object ....Computational infrastructure for developing deep machine learning models. Computational infrastructure for developing deep machine learning models:
The computational infrastructure for developing deep machine learning models aims to enable new developments in machine learning of deep neural network models by providing the specialised computing necessary to train and evaluate the networks. In the last three years, deep networks have smashed previous performance ceilings for tasks such as object recognition in images, speech recognition and automatic translation, bringing the prospect of machine intelligence closer than ever. Modern machine learning techniques have had huge impact in the last decade in fields such as robotics, computer vision and data analytics. The facility would enable Australian researchers to develop, learn and apply deep networks to problems of national importance in robotic vision and big data analytics. Read moreRead less
Visual sensing for localisation and mapping in mining. The creation of high quality survey data is integral to productivity and safety in mining and mining exploration. The current state-of-the-art mine surveying involves scanning from a number of fixed points using laser range-finding equipment (LIDAR). The aim of this project is to develop camera systems and computer vision algorithms to improve the speed and accuracy of this digital mapping of mines, to allow accurate mapping in locations den ....Visual sensing for localisation and mapping in mining. The creation of high quality survey data is integral to productivity and safety in mining and mining exploration. The current state-of-the-art mine surveying involves scanning from a number of fixed points using laser range-finding equipment (LIDAR). The aim of this project is to develop camera systems and computer vision algorithms to improve the speed and accuracy of this digital mapping of mines, to allow accurate mapping in locations denied GPS, and in locations where LIDAR cannot be deployed. The project aims to develop methods to assess these data to detect long-term trends such as shifts in mine drives which may be indicative of stress build-up. The new technology intends to impact both productivity and safety within mining.Read moreRead less
Development Of An Artificial Vision System For Motion Detection. This work will have a broader impact beyond the advancement of a particular area of neurophysiology since it aims to produce a simple, robust velocity estimator in computer models and subsequently in silicon chips. We have already demonstrated the feasibility of transferring our basic preliminary models onto VLSI chips. Consistent with National Research Priorities 3 (Frontier Technologies), 4 (Defence Technologies) & 2 (Promoting a ....Development Of An Artificial Vision System For Motion Detection. This work will have a broader impact beyond the advancement of a particular area of neurophysiology since it aims to produce a simple, robust velocity estimator in computer models and subsequently in silicon chips. We have already demonstrated the feasibility of transferring our basic preliminary models onto VLSI chips. Consistent with National Research Priorities 3 (Frontier Technologies), 4 (Defence Technologies) & 2 (Promoting and Maintaining Good Health) our work has applications in the area of miniature autonomous systems in defence and civilian roles, including miniature unmanned aerial vehicles and collision avoidance detectors, while in the longer term it will serve as the basis for designing implantable artificial eyes.Read moreRead less
Deep reinforcement learning for discovering and visualising biomarkers. This project aims to develop novel methods for discovering and visualising optimal bio-markers from chest computed tomography images based on extensions of recently developed deep reinforcement learning techniques. The extensions proposed in this project will advance medical image analysis by allowing an efficient analysis of large dimensionality inputs in their original high resolution. In addition, this project will be the ....Deep reinforcement learning for discovering and visualising biomarkers. This project aims to develop novel methods for discovering and visualising optimal bio-markers from chest computed tomography images based on extensions of recently developed deep reinforcement learning techniques. The extensions proposed in this project will advance medical image analysis by allowing an efficient analysis of large dimensionality inputs in their original high resolution. In addition, this project will be the first approach capable of discovering previously unknown biomarkers associated with important clinical outcomes. The project will validate the approach on a real-world case study data set concerning the prediction of five-year survival of chronic disease.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.