Driving Scene Understanding – State of The Art and Open Challenges with A/Prof Kourosh Khoshelham скачать в хорошем качестве

Driving Scene Understanding – State of The Art and Open Challenges with A/Prof Kourosh Khoshelham

UNSW AI Institute and UNSW Canberra jointly present: The Long Road Ahead Seminar 2: “Driving Scene Understanding – State of The Art and Open Challenges” with A/Prof Kourosh Khoshelham Abstract Autonomous driving is one of the most exciting innovations of our time. Autonomous vehicles improve the safety of road transport and provide mobility options for elderly and impaired users. In some applications autonomous vehicles carry out the driving tasks that human drivers find tedious and unpleasant. To navigate safely and efficiently, an autonomous vehicle should have the ability to sense the road environment using onboard sensors and recognise the traffic events and participants. However, the complexity, dynamism and geographical diversity of road environments make automated scene understanding a major challenge. In this talk I will provide an overview of the state of the art in driving scene understanding and discuss recent developments and trends in computer vision and machine learning for autonomous vehicles. I will also discuss some of the open challenges in driving scene understanding which must be overcome before autonomous vehicles can hit the road. About the Speaker Kourosh Khoshelham is an associate professor in spatial information at the University of Melbourne. He received his PhD degree in Geoinformatics from Hong Kong Polytechnic University in 2004 and held assistant professor positions at Delft University of Technology and University of Twente before joining the University of Melbourne in 2015. He has co-authored over 170 refereed publications on topics related to spatial information and 3D computer vision. His research interests include autonomous mapping, positioning and navigation, augmented reality, and development of machine learning methods for automated interpretation of imagery and point clouds. Kourosh is the recipient of several prestigious awards including the ISPRS President's Honorary Citation for special, personal, and meritorious contributions to the International Society for Photogrammetry and Remote Sensing. The Long Road Ahead Seminar Series - Self-driving vehicles will bring capillary and irreversible transformations into the worlds of transportation, logistics and urban design. However, it is now clear that the development of this AI-based technology constitutes a long-term complex challenge: despite some impressive technological achievements and the boastful announcements often made by makers and entrepreneurs, autonomous vehicles are not roaming our roads yet and it is still unclear when they will eventually be mature enough to operate in ‘full’ or ‘semi’ autonomous mode. What is still needed to make this innovation real? What are the problems faced by developers and makers, and do these problems have a communal root? We will tackle these questions with the help of authoritative experts from academia, private sector, and public administration. Our seminar aims to identify and investigate with an integrative and interdisciplinary approach the fundamental problems faced by the sector. Please contact [email protected] for any inquires regarding how to host an UNSW.ai sponsored workshop. UNSW AI Institute is the flagship research institute of UNSW (which is one of the World’s Top 100 Universities and 23rd globally for Engineering and Technology*). Its researchers’ extensive track record in AI research and development capabilities across several faculties is well recognised globally. Website: UNSW.ai Twitter:  @unsw_ai LinkedIn:  linkedin.com/company/unsw-ai-institute Youtube: http://unsw.to/youtube_ai