Towards Large-Scale Data from Real-World Environments

During the last 30 years, human-centered perceptual understanding tasks such as face/body detection/ tracking, facial characteristic analysis (e.g., age, gender, ethnicity, and expression), face/gait based human identification, and from understanding simple hand gesture to complex human body language, activity and behaviour, have attracted researchers from both computational and psychological communities. To be able to build tractable models has significant scientific and practical values. For example, human-computer interaction, video indexing, visual surveillance, and Internet vision are active application areas. To that end, researchers have made substantial progress. However, significant challenges still remain.

This one-day workshop (AMFG2010) will provide a focused international forum to bring together researchers and research groups to review the status of recognition, analysis and modeling of face, gesture, and body, to discuss the challenges that we are facing, and to explore future directions. A particular interest and focus of this workshop will be placed upon large-scale data from real-world environments. The workshop will consist of two to three invited talks together with peer-reviewed regular papers (oral and poster). Original high-quality papers are solicited on a wide range of topics including:

* Face and body language recognition in real-world environments with unknown/changing background, e.g., Internet image search, or Facebook, or crowded public spaces.
* Soft biometrics and profiling: age, gender, and ethnicity classification by facial and/or body descriptions.
* Novel mathematical modelling and algorithms, sensors and modalities for face & body gesture/action representation, analysis and recognition.
* Novel applications based on robust detection, tracking and recognition of face, body and action.
* Psychological studies that can assist in understanding computational modelling and building better automated computer vision systems.
* Motion analysis, tracking and extraction of face and body models from image sequences captured by a single or multiple cameras from either a fixed or mobile platform.
* Detection and recognition of face and body under large 3D rotation, illumination change, partial occlusion, and aging.
* Face, gait and action recognition in low-quality, low-resolution video from fixed or mobile cameras.
* Dynamics and learning for gesture, activity, and body behavior interpretation.

Workshop Co-Chairs

* Wen-Yi Zhao, Intuitive Surgical, Inc.
* Xiaoou Tang, Chinese University of Hong Kong
* Shaogang Gong, Queen Mary University of London
* Shuicheng Yan, National University of Singapore