It is our great pleasure to welcome all of you to the 4thACM/IEEE Workshop on Analysis and Retrieval of Tracked Events and Motion in Imagery Streams -- ACM/IEEE ARTEMIS 2013 which is being held this year in Barcelona, Spain, in conjunction with ACM Multimedia. The mission of this workshop is to present the current research advantages in the area of cognitive video supervision and analysis of events, actions and workflows, a critical research task for many reallife multimedia applications. ACM/IEEE ARTEMIS 2013 gives researchers a unique opportunity to share their perspectives with their colleagues interested in the various aspects of video supervision and event analysis.
The call for papers attracted submissions from all over the world. Ten articles were finally accepted after a peer review process. Each paper was judged by at least three reviewers, but most of them by four or five reviewers. These papers have been organized in three sessions. More specifically, the first session is dedicated to "Video Features and Scene Analysis" and includes algorithms, techniques and methods for video scene analysis. In particular, the first article deals with the application of a new algorithm that modifies optical flow on the use of textural features. Then, the next article introduces methods for hand gesture recognition on exploiting depth sensors from KinectTM cameras. The third paper describes a foreground detection method for traffic scenes within complex real-world urban environments. Finally, the fourth article of this session discusses the introduction of transfer learning for re-identifying persons from multiple cameras.
The second session deals with "Retrieval of Multimedia Objects/Events". This session includes two papers. The first proposes a new non-parametric method for clustering image content with mrespect to the semantics. The clustering exploits visual features like the SIFT transform while cultural heritage objects are surveyed. The second and last paper of this session describes a methodology for synchronizing two video sources which capture the same scene from different views.
The third session describes events detection and abnormal behavior recognition from complex visual data. The title of this session is "Analysis of Visual Events". We start with an article which applies methods for abnormal behavior recognition from complex visual scenes. Then, we propose techniques for recognizing complex behaviors based on a human constrained descriptor and adaptable neural networks. The third article applies computer vision tools for maritime detection in outdoor environments of sea ports. Finally, the last but not least paper of the workshop describes a methodology for cross-domain traffic scene understanding by the use of motion models.
All these sessions are presented in a single track one day workshop. The heart of this effort is the researchers who have provided the content of this event. The role of the Program Committee and of the external reviewers was prominent since they worked hard in organizing the workshop, reviewing the papers and providing suggestions for their improvements.
This event is supported by the European funded projects, EXPERIMEDIA, "Experiments in live social and networked MEDIA experiences" (www. www.experimedia.eu/ ), 4D-CH-World, "Four Dimensional Cultural Heritage World" (www.4d-ch-world.eu), and the Greek-Cypriot Collaborative project POSEIDON "Development of an Intelligent System for Coast Monitoring using Camera Arrays and Sensor Networks" (http://www.poseidonproject.eu/) and e-Park, "Exploitation of New Technological Trends in Management and Payment of Public Parking," (www.e-park.eu). In addition the Greek National Projects Viopolis and i-Promotion support this event.
Proceeding Downloads
On improving the robustness of variational optical flow against illumination changes
The brightness constancy assumption is the base of estimating the flow fields in most differential optical flow approaches. However, the brightness constancy constraint easily violates with any variation in the lighting conditions in the scene. Thus, ...
Hand gesture recognition with depth data
Depth data acquired by current low-cost real-time depth cameras provide a very informative description of the hand pose, that can be effectively exploited for gesture recognition purposes. This paper introduces a novel hand gesture recognition scheme ...
Nobody likes Mondays: foreground detection and behavioral patterns analysis in complex urban scenes
Streams of images from large numbers of surveillance webcams are available via the web. The continuous monitoring of activities at different locations provides a great opportunity for research on the use of vision systems for detecting actors, objects, ...
Domain transfer for person re-identification
Automatic person re-identification in is a crucial capability underpinning many applications in public space video surveillance. It is challenging due to intra-class variation in person appearance when observed in different views, together with limited ...
A non-parametric unsupervised approach for content based image retrieval and clustering
Nowadays, there are available extremely large collections of images located on distributed and heterogeneous platforms over the web. The proliferation of billions of shared photos has outpaced the current technology for browsing such collections, but at ...
Warping trajectories for video synchronization
Temporal synchronization of multiple video recordings of the same dynamic event is a critical task in many computer vision applications e.g. novel view synthesis and 3D reconstruction. Typically this information is implied, since recordings are made ...
Abnormal crowd behavior detection and localization using maximum sub-sequence search
This paper presents a novel framework for anomaly event detection and localization in crowded scenes. We propose an anomaly detector that extends the Bayes classifier from multi-class to one-class classification to characterize normal events. We also ...
Behavior recognition from video based on human constrained descriptor and adaptable neural networks
In this paper we introduce a new descriptor, the Human Constrained Pixel Change History (HC-PCH), which is based on Pixel Change History (PCH) but focuses on the human body movements over time. We propose a modification of the conventional PCH which ...
Background modeling methods for visual detection of maritime targets
We propose a system for real-time detection of maritime targets based on monocular video data. In the absence of a priori knowledge about their appearance, targets are detected implicitly via the statistical modeling of the scene's nonstationary ...
Cross-domain traffic scene understanding by motion model transfer
This paper proposes a novel framework for cross-domain traffic scene understanding. Existing learning-based outdoor wide-area scene interpretation models suffer from requiring long term data collection in order to acquire statistically sufficient model ...
- Proceedings of the 4th ACM/IEEE international workshop on Analysis and retrieval of tracked events and motion in imagery stream