A traditional remote sensing system deploy sensors in designated areas and uses a human operator to analyse and interpret sensor data. We propose to reverse the role of man and machine in such situations– augmenting a sensor network with crowdsourcing and using convolutional neural networks (CNN) to automatically analyse sensor data. The talk will present a user modelling system that can be used to develop application and device independent personalized user interfaces to collect data from victims of a disaster zone. We shall also present a modular drone payload that can automatically identify disaster victims from a drone camera using a CNN.Link to Crowd Sourcing Paper:
Link to Geo Tagging Portal:
Музыка
Фильмы
ТВ Шоу
Категории видео
