1st IEEE Internation Workshop on Computer Vision in Sports (at CVPR 2013)

Motivation
Sports is said to be the social glue of society. It allows people to interact irrespective of their social status, age etc. With the rise of the mass media a significant quantity of resources have been channeled into sports in order to improve understanding, performance and presentation. For example, areas like performance assessment, which were previously mainly of interest to coaches and sports scientists are now finding applications in broadcast and other media, driven by the increasing use of on-line sports viewing which provides a way of making all sorts of performance statistics available to viewers. Computer vision has recently started to play an important role in sports as seen in for example football where computer vision-based graphics in real-time enhances different aspects of the game.

Computer vision algorithms have a huge potential in many aspects of sports ranging from automatic annotation of broadcasted footage, through to better understand of sport injuries, and enhanced viewing. So far the use of computer vision in sports has been scattered between different disciplines.


Program (June 28):

9.15 – 9.25: Welcome
9.25 – 10.15: Keynote 1: Computer Vision for sports coverage on television. Graham Thomas, BBC (abstract below)
10.15 – 10.45: BREAK
10.45 – 12.00: Oral session 1

  • Recognising Team Activities from Noisy Data. Alina Bialkowski, Patrick Lucey, Peter Carr, Simon Denman, Iain Matthews and Sridha Sridharan
  • Automatic Recognition of Offensive Team Formation in American Football Plays. Indriyati Atmosukarto, Bernard Ghanem, Karthik Muthuswamy and Narendra Ahuja
  • Sports Type Classification using Signature Heatmaps. Rikke Gade and Thomas Moeslund

12.00 – 13.30: LUNCH
13.30 – 14.10: Keynote 2: Actions in the Eye: From Hollywood to Sports. Cristian Sminchisescu, Lund University (abstract below)
14.10 – 15.25: Oral session 2

  • Visible-Spectrum Gaze Tracking for Sports. Bernardo Pires, Myung Hwangbo, Michael Devyver and Takeo Kanade
  • Non-Invasive Soccer Goal Line Technology: A Real Case Study. Paolo Spagnolo, Pier Luigi Mazzeo, Marco Leo, Massimiliano Nitti, Ettore Stella and Arcangelo Distante
  • Reconstruction of 3D Trajectory for Performance Analysis in Table Tennis. Sho Tamaki and Hideo Saito

15.25 – 15.55: BREAK
15.55 – 17.35: Oral session 3

  • Real-time Person Detection and Tracking in Panoramic Video. Werner Bailer and Marcus Thaler
  • Object Tracking by Occlusion Detection via Structured Sparse Learning. Tianzhu Zhang, Bernard Ghanem, Changsheng Xu and  Narendra Ahuja
  • A Scale and Rotation Invariant Approach to Tracking Human Body Parts Regions in Videos. Yihang Bo and Hao Jiang
  • Athlete Pose Estimation from Monocular TV Sports Footage. Mykyta Fastovets, Jean-Yves Guillemaut and Adrian Hilton

17.35 – 17.50: Closing remarks

 


Keynotes

  • Speaker: Graham Thomas, BBC, UK
    Title: Computer Vision for sports coverage on television
    Abstract:
    Sports coverage on TV can be significantly enhanced by the use of computer vision and 3D graphics, to help analyse an event so that the programme presenters can give a clear and visually-compelling presentation to viewers.  BBC Research & Development has been working closely with BBC Sport for many years to develop new algorithms to meet production requirements, and working with commercial graphics companies to turn these into products.  This talk will tell some of the stories behind this work, including real-time camera tracking, segmentation and 3D reconstruction. The difficulties that have to be overcome in converting an algorithm that works in the lab to something fit for live use in a TV production environment will be discussed.
  • Speaker: Cristian Sminchisescu, Lund University, Sweden
    Title: Actions in the Eye: From Hollywood to Sports
    Abstract:
    Systems based on bag-of-words models operating on image features collected at maxima of sparse interest point operators have been successful for both computer-based visual object and action recognition tasks. While the sparse, interest-point based approach to recognition is not inconsistent with visual processing in biological systems that operate in ”saccade and fixate” regimes, the knowledge, methodology, and emphasis in the human and the computer vision communities remains sharply distinct. Here, we make several contributions aiming to bridge this gap. First, we complement existing state-of-the art large-scale dynamic computer vision datasets like Hollywood-2 and UCF Sports with human eye movements collected under the ecological constraints of the visual action recognition task. To our knowledge these are the first human eye tracking datasets of significant size to be collected for video (497,107 frames, each viewed by 16 subjects), unique in terms of their (a) large scale and computer vision relevance, (b) dynamic, video stimuli, (c) task control, as opposed to free-viewing. Second, we introduce novel dynamic consistency and alignment models, which underline the remarkable stability of patterns of visual search among subjects. Third, we leverage the large amounts of collected data in order to pursue studies and build automatic, end-to-end trainable computer vision systems based on human eye movements. Our studies not only shed light on the differences between computer vision spatio-temporal interest point image sampling strategies and human fixations, as well as their impact for visual recognition performance, but also demonstrate that human fixations can be accurately predicted, and when used in an end-to end automatic system, leveraging some of the most advanced computer vision practice, can lead to state of the art results. The dataset and related papers are available online at: http://vision.imar.ro/eyetracking/.This is joint work at the Institute of Mathematics (IMAR) with Stefan Mathe.

 

Call for papers
The ambition of this workshop is to bring together practitioners and researchers from different disciplines to share ideas and methods on current and future use of computer vision in sports. To this end we welcome computer vision-based research contributions as well as best-practice contributions focusing on the following (and similar) topics:

– estimation of position and motion of cameras and participants in sports

– tracking people and objects in sports

– activity recognition in sports

– event detection in sports

– spectator monitoring

– annotation and indexing in sports

– graphical effects in sports

– analysis of injuries in sports

– performance assessment in sports

 

Important dates

  • Submission deadline: April 7 (extended)
  • Notification of acceptance: April 29
  • Camera ready version: May 10
  • Workshop date: June 28


Submission instructions
Same as for CVPR: Link
Paper submission: Link

 


Program committee:
Ahuja N., Univ. of Illinois at Urbana-Champaign, USA
Chellapa R., Univ. of Maryland, USA
Christmas B., Univ. of Surrey, UK
Cucchiara R., Univ. Modena Reg. Emilia, Italy
Larry Davis, Univ. of Maryland, USA
Dawes R., BBC R&D, UK
Doulamis A.D., Technical University of Crete, Greece
Escalera S., Univ. Of Barcelona, Spain
Ferryman J., Univ. Reading, UK
Gall J., Max Planck Inst., Germany
Ghanem B., ADSC, Singapore
Grau, O, Intel Visual Computing Institute
Guan L., Ryerson Univ., Toronto, Canada
Gong S., Queen Mary Univ. of London, UK
Gonzalez J., UAB- CVC, Catalonia, Spain
Han J., Univ. of Technology Eindhoven, The Netherlands
Hilton A., Univ. of Surrey, UK
Ikizler-cinbis N., Univ. Boston, USA
Kopf S., Univ. of Mannheim, Germany
Maybank S., Birkbeck Univ. of London, UK
Perales F., Univ. de les Illes Balears, Spain
Prati A., Univ. Iuav di Venezia, Italy
Qingming H., Chinese Academy of Sciences, Beijing, China
Sigal L., Disney Reseach, USA
Sminchisescu C., Lund Univ., Sweden

 

Organizers
Thomas Moeslund, Aalborg University, Denmark
Graham Thomas, BBC, UK

 

Overall Meeting Sponsors

  CVF logo      IEEE Computer society