Erik Rosen, Zenuity AB, Sweden. Part 1: , Part 2:
Uwe Franke, Daimler AG, Germany.
Dan Levi, General Motors Research, Israel.
Raquel Urtasun, University of Toronto, Canada
Daniel Cremers, TUM, Germany
Title:Approaching the Perception Challenges of Self-Driving Cars With Robustness and Safety as the Main Priorities
Abstract:Deep learning has revolutionized the self-driving technology space along the last years. In particular, the perception capabilities of autonomous cars have been dramatically boosted by Deep Learning techniques, leading to a much better understanding of traffic scenes from vision, LiDAR and radar information. This talk will briefly introduce Zenuity, a recently launched joint venture between Volvo Cars and Autoliv, and its plans within the autonomous driving space. Furthermore, it will present different examples of the new perception functionalities that Zenuity's Deep Learning and vision teams hav developed since the start of the company, some of them being currently carried over to product development phases. Challenges related to robustness and safety of such perception functionalities will be raised and potential countermeasures discussed.
Title:Towards a Comprehensive Super-Pixel Representation of Traffic Scene Images
Abstract:Most autonomous vehicles on the road fuse information gathered by Radar, Lidar and Vision in order to maximize reliability of the perception system. While early attempts worked on an object level, today medium or even low level fusion is favored. This talk will summarize 10 years research on a compact medium-level representation of traffic scene images named Stixel-World. This representation encodes depth, motion and semantics in only 4kByte, at the same time making the content of the scene explicit. Years of practice have proven its power. I will concentrate on recent progress and compare this representation developed for traffic scene images against alternative super-pixel approaches. The obtained representation is not limited to image data but can also be efficiently used to encode geometry and semantics of Lidar or Lidar-Camera combinations.
Title:Training models for road scene understanding with automated ground truth
Abstract:Collecting and labeling training data for vision-based road scene understanding is a major challenge. The most prominent approach is to use manual labeling, though it is clear that scalability of this approach is limited. More scalable alternatives are simulated data and cross-sensor label transfer. In this talk I will present automatically generated ground truth using one or more sensors, primarily dense Lidar. Specifically, I will present the benefits and challenges of this approach for road scene understanding tasks, including general and category-based obstacle detection, free space and curb detection.
Title:Direct Methods for 3D Reconstruction & Visual SLAM with Applications to Autonomous Systems
Abstract:The reconstruction of the 3D world from images is among the central challenges in computer vision. Starting in the 2000s, researchers have pioneered algorithms which can reconstruct camera motion and sparse feature-points in real-time. In my talk, I will introduce direct methods for camera tracking and 3D reconstruction which do not require feature point estimation, which exploit all available input data and which recover dense or semi-dense geometry rather than sparse point clouds. Applications include 3D free-viewpoint television and autonomous vehicles.
Topics of Interest
Analyzing road scenes using cameras could have a crucial impact in many domains, such as autonomous driving, advanced driver assistance systems (ADAS), personal navigation, mapping of large scale environments and road maintenance. For instance, vehicle infrastructure, signage, and rules of the road have been designed to be interpreted fully by visual inspection. As the field of computer vision becomes increasingly mature, practical solutions to many of these tasks are now within reach. Nonetheless, there still seems to exist a wide gap between what is needed by the automotive industry and what is currently possible using computer vision techniques. The goal of this workshop is to allow researchers in the fields of road scene understanding and autonomous driving to present their progress and discuss novel ideas that will shape the future of this area. In particular, we would like this workshop to bridge the large gap between the community that develops novel theoretical approaches for road scene understanding and the community that builds working real-life systems performing in real-world conditions. To this end, we encourage submissions of original and unpublished work in the area of vision-based road scene understanding. The topics of interest include (but are not limited to):
- Road scene understanding in mature and emerging markets
- Deep learning for road scene understanding
- Prediction and modeling of road scenes and scenarios
- Semantic labeling, object detection and recognition in road scenes
- Dynamic 3D reconstruction, SLAM and ego-motion estimation
- Visual feature extraction, classification and tracking
- Design and development of robust and real-time architectures
- Use of emerging sensors (e.g., multispectral, RGB-D, LIDAR and LADAR)
- Fusion of RGB imagery with other sensing modalities
- Interdisciplinary contributions across computer vision, optics, robotics and other related fields.
We encourage researchers to submit not only theoretical contributions, but also work more focused on applications. Each paper will receive 3 double blind reviews, which will be moderated by the workshop chairs.<--
- Erik Rosen, Zenuity AB, Sweden
- Uwe Franke, Daimler AG, Germany
- Dan Levi, General Motors Research, Israel
- Raquel Urtasun, University of Toronto, Canada
- Daniel Cremers, TUM, Germany
- Submission Deadline: EXTENDED ! July 25th (23:59 Pacific Time).
- Notification of Acceptance: August 18th.
- Camera-ready Deadline: August 25th.
- Workshop: October 23rd.
- Prof. Fredrik Kahl, Chalmers University of Technology
- Dr. Bart Nabbe (Aurora Innovation, USA)
- Dr. Mathieu Salzmann (EPFL, Switzerland)
- Dr. Lars Petersson (data61 CSIRO, Australia)
- Dr. Jose Alvarez (data61 CSIRO, Australia)
- Agata Mosinska, EPFL
- Albert Jimenez, UPC
- Alexander Schwing, UIUC
- Ali Armin, Data61
- Anders Eriksson, QUT
- Andrii Maksai, EPFL
- Anurag Arnab, Oxford University
- Artem Rozantsev, EPFL
- Basura Fernando, Australian National University
- Bugra Tekin, EPFL
- Carl Toft, Chalmers
- Carl Olsson, Lund University, Sweden
- Carl Henrik Henrik, Bristol University
- Carlos Becker, Pix4D
- Carlos Fernandez,Universidad de Alcala
- Chuong Nguyen, Data61
- David Bermudez, Computer Vision Center
- Eduard Trulls, EPFL
- Erik Bylow, Lund University
- Fatemeh Saleh, Data61
- Helge Rhodin, EPFL
- Henrik Aanæs, Technical University of Denmark
- HONGDONG LI, Australian National University
- Kneip Laurent, Australian National University
- Ksenia Konyushova, EPFL
- Kwang Yi, EPFL
- Lachlan Tychsen-Smith, Data61
- Magnus Oskarsson, Lund University
- Måns Larsson, Chalmers
- Markus Enzweiler, Daimler R&D
- Mårten Wadenbäck, Lund University
- Mehrtash Harandi, Data61
- Miaomiao Liu, Data61 - CSIRO
- Mohammad Najafi, Oxford University
- Pablo Márquez-Neila, EPFL
- Pierre Baque, EPFL
- Roger Bermudez Chacon, EPFL
- Sadeep Jayasumana, Oxford University
- Sadegh Aliakbarian, Data61
- Salman Khan, Data61
- Sarah Namin, ANU
- Shaodi You, Data61-CSIRO
- Shuai Zheng, Oxford University
- Simon Jegou, MILA
- Thomas Probst, ETH Zurich
- Timo Rehfeld, Daimler
- Timur Bagautdinov, EPFL
- Xuming He, Australian national University
- Yuhang Zhang, Chalmers
- Zeeshan Hayder, Data61
Papers should describe original and unpublished work about the above or closely related topics. Each paper will receive double blind reviews, moderated by the workshop chairs. Authors should take into account the following:
- All papers must be written in English and submitted in PDF format.
- Papers must be submitted online through the CMT submission system. The submission site is: https://cmt3.research.microsoft.com/CVRSUAD2017.
- The maximum paper length is 8 pages (including references). Note that shorter submissions are also welcome. The workshop paper format guidelines are the same as the Main Conference papers.
- Submissions will be rejected without review if they: contain more than 8 pages, violate the double-blind policy or violate the dual-submission policy. The author kit provides a LaTeX2e template for submissions, and an example paper to demonstrate the format. Please refer to this example for detailed formatting instructions.
- A paper ID will be allocated to you during submission. Please replace the asterisks in the example paper with your paper's own ID before uploading your file. More detailed instructions can be found at the main conference website.