Agenda

Categoría: Empleo

Descripción: The PhD is focused on understanding what is happening in a video-surveyed scene of interest, because of its critical or security nature, including both activities and behaviour, in the case of individuals, and anomalous events in the case of objects. This demand for intelligent monitoring occurs both indoors and outdoors, in realistic situations (in the wild), including: hot spots in cities, stadiums gateways and surroundings or facilities for leisure and recreation, railway lines, roads, level crossings, access to transportation (train, plane, etc.), public buildings, shopping malls, and, in general, any place where an abnormality detection is relevant to safety.
The objective of the PhD is to contribute to the state-of-the-art in automatically process video sequences of different scenarios of interest and observe the typical and atypical (usual or unusual) activities that take place there in order to determine whether it is a normal or abnormal situation and to monitor in real-time the infrastructure and its usage mainly for safety reasons, in order to decide, at any time, if there is an urgent action to take.
In this context, the PhD will propose new strategies for the detection of group and individual behaviour to identify anomalies applying innovative computer vision and audio, and machine learning techniques, as well as definition and techniques for detecting/obtaining and integrating physical and semantic space-time audio-visual attributes characteristics related to human appearance or behaviour.
Deep learning is widely recognized as one active research direction in machine learning. Its performance already reaches that of human beings in many real learning scenarios, including computer vision. Based on its high interest, the objective of the intended project is to investigate the state-of-the-art deep learning models, testing well-known and novel architectures and algorithms of deep neural networks to design the aforementioned system. Further enhanced technologies based in deep-learning are expected to benefit public security, among other social objectives.
The intended project is not only theoretically significant, but more importantly the research outcomes could be extensively applied in various applications, thus the construction of realistic demonstrators and training databases are also key objectives of the proposal.
The PhD is involved in HEIMDAL project (TIN2016-75982-C2-1-R), sponsored by the Spanish government.

Fecha Inicio: 21-09-2017

Fecha Fin: 16-11-2017

Más Información: http://www.geintra-uah.org

Convocante: GEINTRA Group

URL: http://www.geintra-uah.org

Fichero Adjunto:

• 20170928-PhDCall.pdf