On April 16th 2015 a meeting was organised in Padua by Slándáil partners to explore the relationship between speech and image analytics. The project has found that emergency communications can be expedited by good video presentation or sometimes hampered by by bad presentation. Ongoing work on Speech Analytics by the University of Padua (led by Maria Grazia Busa) is analysing aspects such as pitch change and gestural deixis (i.e. words and phrases that cannot be fully understood without additional contextual information).
Ulster University (represented by Bryan Scotney) took part in the meeting with a focus on identifying named entities and movements made by tracked entities in image analysis. Automatic gestural deixis, involving person, temporal and place, should help to improve multimedia communications during emergencies.
Above: A comparison between gestural communication used as an example in the meeting slides. Slides are available below.
This collaboration between the two tasks has important benefits for Multi-modal aggregation at Trinity College Dublin (led by Khurshid Ahmad).
A discussion took place between Maria Grazia Busà, Bryan Scotney, Khurshid Ahmad, Maria Teresa Musacchio, Raffaella Panizzon, Sara Brugnerotto. The Lead Beneficiaries concluded that image analytics will deal with gestures as a target area and a provision has to be made of training videos in this area.
A further workshop is planned in Padova as a part of a meeting organised by University of Padova in May 2015.