Call for Papers

Visual surveillance is a major research area in computer vision. The scientific challenge in crowd image analysis is to devise and implement automatic systems for obtaining detailed information about the movements of, and collective behaviour of, individuals and groups within a crowded scene observed by a single camera or by a network of cameras.  

Crowd image analysis is a key technology in the following areas:

• the fight against terrorism and crime including threat assessment
• public safety, especially in transport networks, town centres and public facilities such as schools, hospitals and sports grounds
• the efficient management of transport networks and public facilities

The growth in the development of the field has not been met with complementary systematic performance evaluation of developed techniques. It is especially difficult to make comparisons between algorithms if they have been tested on different datasets under widely varying conditions. 

A one day workshop is being held in Miami in conjunction with CVPR 2009. The workshop continues the theme of the highly successful PETS workshops. This workshop aims to bring together researchers interested in performance evaluation of visual tracking and surveillance algorithms. The workshop is also unique in that all participants are evaluating algorithms on the same datasets. Further to this, the workshop is an opportunity to present and discuss methodologies and criteria for objective evaluation of visual surveillance algorithms.  

For PETS 2009, the theme is multi-sensor tracking and event recognition in crowded public areas. This includes low-level analysis: crowd person count and density estimation; mid-level analysis: tracking of an individual or individuals within a crowd; high-level analysis: detection of separate flows and specific crowd events.

Submissions are solicited which: 

• Contribute to performance evaluation methodologies for tracking and/or surveillance. This can be a general paper outlining, for example, a new methodology for evaluation, or performance evaluation best practice etc. It is not necessary to explictly consider the datasets provided for the workshop, however you are encouraged to use the PETS datasets available here.

OR

• Describe an approach to event recognition and report results based on the datasets provided for this workshop. The paper should include evaluation of the results obtained on the data (to be submitted in XML format).

The datasets will be made available to participants of the workshop session. The focus of the datasets is on crowd image analysis and includes the following categories: i) person count and density estimation, ii) tracking of individual(s) within a crowd, and iii) anomalous crowd movements, of increasing complexity, captured using multiple sensors.

Important Dates

6 March 2009

Public release of challenge video datasets

30 March 2009

Publication of submission requirements/dd>

7 May 2009

*Extended* deadline for electronic paper submission

26 May 2009

Notification of acceptance

3 June 2009

Deadline for camera-ready version

25 June 2009

Date of workshop in Miami