Tuesday, March 4th, 2025
Full program here
Full program here
Half Day – Morning (0800-1200) MST, Salon H, JW Marriott Starpass in Tucson, Arizona
Winter sports like skiing, bobsleigh, ice skating, and ice hockey and mountain sports such as climbing, mountaineering, hiking and downhill biking enjoy widespread popularity as year-round pastimes. In efforts to enhance human well-being and foster social connections, various sports federations seek to promote these activities by encouraging greater participation. A primary strategy involves the broadcast of professional competitions, which serves to engage audiences and to attract new enthusiasts. Notably, televised events, particularly in alpine and nordic skiing, collectively draw an impressive viewership of approximately 6 million, thereby reaching a substantial audience of potential participants. Similarly, mountain sports, like climbing have seen a consistent growth in the past years, with the inclusion in the Summer Olympic Games in Tokyo 2020 and Oscar-winning documentary ``Free Solo'' contributing to attract new practitioners and viewers to the events.
In order to increase the interest in winter and mountain sports disciplines, the enhancement of the spectator experience is a key element, enriching the broadcasted content with additional content, leveraging the most recent advances in video technologies. While current methods for analyzing athlete performance often rely on sensor-based solutions like IMUs or GNSS, the study, development, and utilization of image and video analytics remains a relatively underexplored area in the winter sports domain. Despite the abundance of visual data available from broadcast videos, user-generated, or collected using ad-hoc cameras, a number of challenges arise, because of the dynamic, like athletes' high speed or unseen body poses, as well as the challenging conditions, such as harsh snowy and icy environments or vertical walls.
These challenges must be carefully addressed, especially considering the real-time processing demands of broadcasting applications and the decision-making processes involved in training scenarios. Tackling these issues presents unique and stimulating challenges, offering opportunities for significant contributions to the field of computer vision.
The workshop invites paper submissions focusing on the analysis and interpretation of images and videos captured during winter sports and related summer activities such as mountain sports (e.g., mountaineering, downhill biking, climbing). Topics of interest for the workshop include machine learning solutions for video understanding, pose estimation of athletes, evaluation of athlete performance, forecasting performance, injury detection/prevention, crowd and spectator monitoring, augmented/virtual reality for fan engagement, and applications of computer vision/AI to various winter and mountain sports disciplines. Additionally, submissions can address challenges such as understanding images/videos in harsh weather conditions, camera pose estimation in broadcast videos, trajectory reconstruction, winter scene reconstruction, snow/ice measurements, real-time processing algorithms, fusion of image/video data with other sensor data, and the creation of datasets and benchmarks.
Research papers are solicited in, but not limited to, the following topic areas:
Machine learning solutions for video understanding or activity recognition regarding winter and mountain sports
Pose estimation of athletes
Evaluation and measurement of athlete performance
Performance forecasting
Detection/evaluation/prevention of injuries in winter sports with computer vision
Crowd and spectators monitoring
Augmented/virtual reality for winter sports and fan engagement
Applications of computer vision/AI to winter sports (skiing, ice-hockey, ice-skating, biathlon, bobsleigh, luge, curling, etc.)
Image/video understanding in winter/harsh weather conditions
Camera pose estimation in broadcast videos
Video-based trajectory reconstruction and analysis
Winter scene reconstruction from images/videos
Snow/ice measurements and analysis with computer vision
Real-time processing algorithms
Fusion of image/video data and other sensor data
Datasets, benchmarks and annotations of winter sport data
Please see the Submissions and Dates page for details about the submissions.
Prof. John Zelek
University of Waterloo
Prof. Rainer Lienhart
University of Augsburg
More details HERE!
Deadline for submissions: 31 January 2025