第十七届图像图形技术与应用学术会议(IGTA2022），由北京图象图形学学会主办，北京科技大学承办，将于2022年4月23-24日在北京科技大学天工大厦召开。本次大会的主题为“图像图形与认知智能”。大会内容包括特邀报告、论文报告、前沿论坛交流、专题论坛、优博展示，青托沙龙，参观展览等部分。图像图形技术在新时代中所起的作用、成果、发展及深层次应用，是大会主要研究交流探讨的内容，现诚邀国内外研究人员和工程技术人员及广大在读学生积极参会。

特邀报告（一）2022年4月23日 09:40 - 10:30

Prof. Nick Barnes

报告题目：Dense prediction with uncertainty for prosthetic vision and bushfire detection.

报告简介：

Computer Vision can support fine-grained human interaction with an object through dense prediction of the object’s spatial extent. In Bushfire detection from tower-based cameras, emergency service dispatch needs to know the source of smoke (bottom of a smoke plume) accurately to within a few kilometres, from up to tens of kilometres away. In prosthetic vision, a device is surgically implanted to restore a sense of vision to profoundly blind individuals. We present object location information via electrical stimulation of the human vision system based on a stream of images from a head mounted camera for avoiding obstacles in mobility tasks and interacting with table-top objects. Both require accurate visual localization of object edges. For this type of human interaction system, we need to segment objects from their background with certainty and consistency over time. Often the objects required do not have large established pixel-wise labelled datasets. Our approaches use dense prediction with uncertainty. We use class agnostic tasks, specifically, saliency, camouflaged objects and smoke segmentation, and have developed datasets and approaches for accurate pixel-wise prediction with uncertainty estimation. Our approaches can create diverse mappings to model tasks that sometimes have a subjective nature. We present results of fire detection from towers around Canberra in Australia, and from implanted participants in our bionic vision trials.

嘉宾简介：

Nick Barnes is a Professor at the School of Computing at the Australian National University. He started full time at ANU in 2019. At the University of Melbourne, he completed BSc (Hons) and PhD in computer vision for navigation in 1999. He was a visiting researcher at the LIRA Lab, University of Genoa, Italy in 1999, and a tenured lecturer at the University of Melbourne in Computer Science and Software Engineering from 2000 to 2003. From 2003-2016 he was with NICTA, Australia’s most successful ICT Centre of Excellence, where he was a Senior Principal Researcher and Led the Computer Vision Research Group. From 2016-2019 he held appointments both at the School of Engineering at the Australian National University, and as Research Group Leader of Computer Vision at CSIRO’s Data61. He was a founding Lead Investigator on the Bionic Vision Australia Special Research Initiative of the Australian Research Council, leading vision processing. He continues to work with Bionic Vision Technologies, a spinout company of Bionic Vision Australia. He is a Lead Investigator, leading the Camera Tower project as part of the Optus-ANU Research Centre of Excellence on Bushfires. He has authored and co-authored more than 150 publications in refereed journals and conferences.

注册参会

ffdfe870b860ff99a5cc5f69483259cf_640_wx_fmt=png&wxfrom=5&wx_lazy=1&wx_co=1.png