Prof. Dr. Katsushi Ikeuchi
University of Tokyo, Japan
Photometric Analysis for Autonomous Systems
Autonomous systems recognize the outer world through the appearance of an outer environment. Most of the previous research pays attention to the geometric relations obtained from the features contained in the appearance. On the other hand, the appearances observed by autonomous systems can be represented as products of illumination conditions and intrinsic characteristics of the outer world. In this talk, we explore these photometric relations among appearances and how to utilize those relations and intrinsic characteristics for guiding autonomous systems.
Dr. Michael James
Toyota Research, USA
Sensors, Maps, and Prior Knowledge: Perception for Automated Vehicles
Advanced driver safety and automated driving technologies hold the promise for increasing safety, convenience, and independence. Over the past few years, we have seen both a growing number and increasing capability of sensing systems on modern automobiles, and this trend is expected to continue. The resulting systems will assist drivers (driver support or warning systems) and control the vehicle in limited situations. However, increasing the amount of automated control is only possible with improved perceptual abilities, and there are many unsolved questions in this area. In this talk, I present research taking place at Toyota Research Institute North America on methods for improving perceptual abilities of vehicles, by combining immediate perceptual information from on-board sensors with unconventional maps, while leveraging prior knowledge. There are many interesting research opportunities in this domain, which I will attempt to highlight.
Dr. Abdelaziz Khiat
Nissan Research, Japan
Autonomous Driving Technologies and Scene Understanding
To materialize the idea of a vehicle that looks out for you, Nissan has been developing its Autonomous Driving (AD) technologies over the years. In this talk, I will explain the expected role of AD technologies in solving some of the current problems facing society. In the context of city autonomous driving, a vehicle needs to gain a continuous understanding of its surround. I will introduce some building functions of our autonomous vehicles and reason about the problematics of scene understanding. Throughout the presentation, I will show a few examples of the realized functions and highlight the remaining challenges.
Prof. Dr. Simon Lucey
Carnegie Mellon University, USA
How to Rule the World with Your Smart Phone
IMUs have a rich history of being combined with monocular vision for robotic navigation and odometry applications. These IMUs require sophisticated and quite expensive hardware rigs to perform well. IMUs in smart devices, however, are chosen for enhancing interactivity - a task which is more forgiving to noise in the measurements. We demonstrate, however, that the ubiquity of these “noisy” IMUs makes them increasingly useful in modern computer vision algorithms. Indeed, we show in this work how an IMU from a smart device can help a face tracker to measure pupil distance, and an SfM algorithm to measure the metric size of objects. We also identify motions that produce better results, and develop a heuristic for estimating, in real-time, when enough data has been collected for an accurate metric reconstruction.
I will also discuss some of my laboratory's recent industrial engagements, and how metric reconstruction played a role. Of specific interest will be work we have recently completed in the space of obtaining dense semantic 3D face scans using hand held mobile/tablet devices. This technology has the potential to revolutionize many aspects of traditional “bricks and mortar” retail by allowing for the virtual try on of wearable merchandise (e.g. glasses, makeup, etc.) using handheld devices.
Dr. David Pfeiffer
Daimler Research, Germany
Making Bertha Drive
Research and development in the field of intelligent vehicles concentrated on active safety systems like lane-departure protection and emergency braking during the last decade. Recently, first autonomous functions were introduced to the market and current development aims at highly autonomous driving functions for highway driving and autonomous parking for the year 2020.
While these functions seem to be realizable, autonomous driving in cities is still a challenge for research. The talk summarizes the experiences gained from the “Bertha” project where the Mercedes Benz S-Class S 500 Intelligent Drive aka “Bertha” followed the historic Bertha Benz memorial route from Mannheim to Pforzheim, Germany, in fully autonomous manner. The course taken by the autonomous vehicle had a length of 103 km and covered rural roads, 23 small villages and major cities (e.g. downtown Mannheim and Heidelberg). The route posed a large variety of difficult traffic scenarios including intersections with and without traffic lights, roundabouts, and narrow passages with oncoming traffic.