Data captured by wearable cameras (egocentric data) can be used to tackle the important issue of pedestrian safety by providing useful information for both citizens and public authorities. Despite significant work in this area, automated visual analysis of egocentric data remains an open problem. PEDESTRIAN project is funded internally by CYENS and is dedicated to the development of advanced visual wearable technologies for improving urban life, and, more specifically, the walking conditions of pedestrians across the spectrum: children, the elderly, pregnant women, persons with disabilities, etc. The main problems that pedestrians face in south and north parts of Nicosia – including the lack of safe crosswalks, overcrowded and narrow footpaths, barriers to pedestrian movement, and periodic road maintenance – will be automatically detected and recognized using computer vision techniques. Two MRGs from two different research pillars (Visual Sciences and Communications and Artificial Intelligence), BIO-SCENT and SCRAT, will collaborate for the successful implementation of the project.
