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
Robust and scalable change detection in geo-spatial data. A flood of data in the form of text, images and video emanate from a proliferation of sensors. These data are collected but rarely analysed, rendering it meaningless. This project aims to develop new software and techniques to detect changes over time in large scale geographically referenced data (for example photomaps) for use across numerous domains.
Design of adaptive learning visual sensor networks for crowd modelling in high-density and occluded scenarios. Partnering University of Melbourne researchers, with video surveillance experts SenSen, engineering consultants ARUP and the Melbourne Cricket Club, the project addresses research enabling a system-integrating, existing surveillance, infrastructure to model crowd behaviour and exit strategies, providing real-time analysis, prediction and response capabilities for venue managers and emer ....Design of adaptive learning visual sensor networks for crowd modelling in high-density and occluded scenarios. Partnering University of Melbourne researchers, with video surveillance experts SenSen, engineering consultants ARUP and the Melbourne Cricket Club, the project addresses research enabling a system-integrating, existing surveillance, infrastructure to model crowd behaviour and exit strategies, providing real-time analysis, prediction and response capabilities for venue managers and emergency services. This new capability enhances utilisation of security resources to prevent injury and fatalities in evacuation scenarios, applicable to existing venues and influencing the development of new facilities around the country. The project delivers researcher training, global clientele for local technology and a platform for local industry growth.Read moreRead less