ETRAN 2017

U radu su opisani koncepti i arhitektura konvolucijskih neuronskih mreža, čija primena je široko rasprostranjena. Značaj primene ove vrste neuronskih mreža je veliki, te su i istraživanja u ovoj oblasti od posebnog interesa. Postignuti su već značajni rezultati, te su napori oko sistematizacije u oblasti opravdani. U radu su predstavljeni važni rezultati istraživanja u oblasti, predložena je konzistentna terminologija i ukazano na izazove daljih istraživanja. 

Nova Science Publishers 2017

One of the most critical aspects of emergency management is the notification system concerning how to get updated and accurate information from the very first stages of the event and how to notify affected people. The citizens, as the simplest level of participation in an emergency scenario, can act as consumers information consumers (e.g. about alarms, affected areas, instructions for handling critical situation), but also as evidence producers (e.g. via phone, social network). With the aim to develop a deeper understanding of potential benefits from using Linked Data technologies in emergency scenarios, in this Chapter we present our current work on building citizen applications for improved emergency response. The paper introduces first the potential areas of applicability of Linked Data in innovative emergency management solutions and points out to early prototypes in Europe based on crowdsourcing, social networks, mobile devices and Open Data. Next, the software requirements and information flow are analyzed for a data ecosystem that facilitates the engagement of citizens leveraging novel approaches such as participatory sensing approach (knowledge from the wisdom of crowds) and cloud-enabled solutions. In order to demonstrate the feasibility of the approach, a prototype application Traffaccs has been developed using state-of-the-art open-source tools and frameworks: the Massachusetts Institute of Technology Panya framework for implementing the mobile client and Dydra graph database for storing the notification messages. 

ICISTS 2017

Nowadays modern mobile phones have a lot of sensors that can be helpful in various situations, for

instance, to facilitate the collaboration between citizens and public services. This paper explores the use of smart phones for producing rich information about traffic accidents in real time. The goal is to examine the Linked Data approach for building a mobile application to be used by citizens for collecting evidence about an exceptional situation (a traffic accident) and for notifying emergency organizations (e.g. the nearest police station) and the wider community. The benefits expected from using the Linked Data in this scenario are related to interlinking data from different sources, semantically enriching the data and sharing the data in a standardized format. This paper presents preliminary results from the design and implementation of a mobile application for Android-based devices.