Call for Papers 1st Autonomous Vehicle Vision (AVVision’21) Workshop

In conjunction with WACV 2021

The Autonomous Vehicle Vision 2021 (AVVision’21) workshop (webpage: avvision.xyz) aims to bring together industry professionals and academics to brainstorm and exchange ideas on the advancement of visual environment perception for autonomous driving. In this one-day workshop, we will have regular paper presentations and invited speakers to present the state of the art as well as the challenges in autonomous driving. Furthemore, we have prepared several large-scale, synthetic and real-world datasets, which have been annotated by the Hong Kong University of Science and Technology (HKUST), UDI, CalmCar, ATG Robotics, etc. Based on these datasets, three challenges will be hosted to understand the current status of computer vision and machine/deep learning algorithms in solving the visual environment perception problems for autonomous driving: 1) CalmCar MTMC Challenge, 2) HKUST-UDI UDA Challenge, and 3) KITTI Object Detection Challenge.

Keynote Speakers:

●      Andreas Geiger, University of Tübingen

●      Ioannis Pitas, Aristotle University of Thessaloniki

●      Nemanja Djuric, Uber ATG

●      Walterio Mayol-Cuevas, University of Bristol & Amazon

Call for Papers:

With a number of breakthroughs in autonomous system technology over the past decade, the race to commercialize self-driving cars has become fiercer than ever. The integration of advanced sensing, computer vision, signal/image processing, and machine/deep learning into autonomous vehicles enables them to perceive the environment intelligently and navigate safely. Autonomous driving is required to ensure safe, reliable, and efficient automated mobility in complex uncontrolled real-world environments. Various applications range from automated transportation and farming to public safety and environment exploration. Visual perception is a critical component of autonomous driving. Enabling technologies include: a) affordable sensors that can acquire useful data under varying environmental conditions, b) reliable simultaneous localization and mapping, c) machine learning that can effectively handle varying real-world conditions and unforeseen events, as well as “machine-learning friendly” signal processing to enable more effective classification and decision making, d) hardware and software co-design for efficient real-time performance, e) resilient and robust platforms that can withstand adversarial attacks and failures, and f) end-to-end system integration of sensing, computer vision, signal/image processing and machine/deep learning. The AVVision'21 workshop will cover all these topics. Research papers are solicited in, but not limited to, the following topics:

• 3D road/environment reconstruction and understanding;
• Mapping and localization for autonomous cars;
• Semantic/instance driving scene segmentation and semantic mapping;
• Self-supervised/unsupervised visual environment perception;
• Car/pedestrian/object/obstacle detection/tracking and 3D localization;
• Car/license plate/road sign detection and recognition;
• Driver status monitoring and human-car interfaces;
• Deep/machine learning and image analysis for car perception;
• Adversarial domain adaptation for autonomous driving;
• On-board embedded visual perception systems;
• Bio-inspired vision sensing for car perception;
• Real-time deep learning inference.

Author Guidelines:

Authors are encouraged to submit high-quality, original (i.e. not been previously published or accepted for publication in substantially similar form in any peer-reviewed venue including journal, conference or workshop) research.

The paper template is identical to the WACV2020 main conference. The author toolkit (latex only) is available both on Overleaf and in Github. The submissions are handled through the CMT submission website: https://cmt3.research.microsoft.com/AVV2021/.

Papers presented at the WACV workshops will be published as part of the “WACV Workshops Proceedings” and should, therefore, follow the same presentation guideliness as the main conference. Workshop papers will be included in IEEE Xplore, but will be indexed separatelly from the main conference papers.

For questions/remarks regarding the submission e-mail: avv.workshop@gmail.com.

Challenges:

Challenge 1: CalmCar MTMC Challenge

Multi-target multi-camera (MTMC) tracking systems can automatically track multiple vehicles using an array of cameras. In this challenge, participants are required to design robust MTMC algorithms, which are targeted at vehicles, where the same vehicles captured by different cameras possess the same tracking IDs. The competitors will have access to four large-scale training datasets, each of which includes around 1200 annotated RGB images, where the labels cover the types of vehicles, tracking IDs and 2D bounding boxes. Identification precision (IDP) and identification recall (IDR) will be used as metrics to evaluate the performance of the implemented algorithms. The competitors are required to submit their pretrained models as well as the corresponding docker image files via the CMT submission system for algorithm evaluation (in terms of both speed and accuracy). The winner of the competition will receive a monetary prize (US$5000) and will give a keynote presentation at the workshop.

Challenge 2: HKUST-UDI UDA Challenge

Deep neural networks excel at learning from large amounts of data but they can be inefficient when it comes to generalizing and applying learned knowledge to new datasets or environments. In this competition, participants need to develop an unsupervised domain adaptation (UDA) framework which can allow a model trained on a large synthetic dataset to generalize to real-world imagery. The tasks in this competition include: 1) UDA for monocular depth prediction and 2) UDA for semantic driving-scene segmentation. The competitors will have access to Ready to Drive (R2D) dataset, which is a large-scale synthetic driving scene dataset collected under different weather/illumination conditions using the Carla Simulator. In addition, competitors will also have access to a small amount of real-world data. The mean absolute value of the relative (mAbsRel) error and the mean intersection over union (mIoU) score will be used as metrics to evaluate the performance of UDA for monocular depth prediction and UDA for semantic driving scene segmentation, respectively. The competitors will be required to submit their pretrained models and docker image files via the CMT submission system.

Challenge 3: KITTI Object Detection Challenge

Researchers of top-ranked object detection algorithms submitted to the KITTI Object Detection Benchmarks will have the opportunity to present their work at AVVision'21, subject to space availability and approval by the workshop organizers. It should be noted that only the algorithms submitted before 12/20/2020 are eligible for presentation at AVVision'21.

Important Dates:

Full Paper Submission: 11/02/2020

Notification of Acceptance: 11/23/2020

Camera-Ready Paper Due: 11/30/2020

HKUST-UDI UDA Challenge abstract and code submission: 12/13/2020

Notification of HKUST-UDI UDA Challenge results: 12/20/2020

CalmCar MTMC Challenge abstract and code submission: 12/13/2020

Notification of CalmCar MTMC Challenge results: 12/20/2020