Afra Mashhadi, Alcatel-Lucent Bell Labs
Afra Mashhadi is a research scientist in the domain of ubiquitous computing in Bell Laboratories. In her previous appointment, she was a post-doctoral researcher first at UCL and then in Bell Labs. Her research interest spans from opportunistic and sensor networks to HCI. Her current research focuses on understanding users participation, accuracy and coverage of ubiquitous crowd-sourcing systems where users voluntarily contribute real-time information about spatial objects/events.
A perfect partnership: Activity Recognition and the Crowd
Recognising and modelling human activity can be applied to a diverse range of real-world applications, from intelligent transportation where one can infer delays occurring in metro line by monitoring the accelerometer of travellers smartphones, to surveillance systems that can detect anomalies and/or problematic behaviours in different neighbourhood. However, taking these models out of the lab context and deploying them into real-world scenarios is not an easy task, as they often require a priori training that is highly dependent on the particular context of the activity. A possible solution is to leverage crowd-sourcing systems so to train the activity recognition on the fly based on the input of the crowd. In this talk we discuss the pros and cons of leveraging the crowd-sourcing techniques in partnership with AR models and present the challenges that need to be faced in this domain.