Where: Sydney, Australia
This workshop addresses the topic of wearable computer vision systems, understanding by that cameras that are worn by a person, somehow mounted or attached to a human body, but not necessarily with the same point of view than the user. The workshop will consist on both invited talks by experts in the field covering different related topics and contributed works by workshop participants.
Thanks to the advances on vision technologies, nowadays cameras are ubiquitous sensors whose presence is constantly growing and offer more and more solutions to real-life problems. Besides, the miniaturization of the camera optics and electronics has facilitated the construction of wearable wide-field-of-view visual sensors. The ever-growing computing capabilities, either locally or in the cloud, are pushing even further the potential for this technology.
Related venues have already been held with great success in previous computer vision conferences. This workshop will continue to gather together the growing community interested in wearable systems based on visual sensors, which is gaining momentum with the release of Google Glass, and the easy access to other commercial wearable cameras, such as Memoto or GoPro.
As part of the workshop, we are arranging and releasing a set of sequences from daily activities and actions, acquired from multiple wearable devices, to allow researchers to try out their algorithms in this kind of data. Devices used include catadioptric, RGB-d and conventional vision systems mounted either on the front, a helmet or the chest, together with inertial measurements. The scenarios cover different trajectories while performing different actions, recorded with several users and acquiring all the sensors simultaneously. The variety of vision systems intends to stimulate the discussion on the possible configurations/tasks for wearable vision systems, since we are covering the same scenarios with all of them.