IMEXT: A method and system to extract geolocated images from Tweets — Analysis of a case study

Abstract

Extracting useful information from social networks raises several challenges that still represent open research issues. In this paper we focus on the problem of extracting geolocated images from Tweets to support emergency response. A Tweet analysis process is discussed, focusing on the selection of posts, their geolocation based on their text content, and the subsequent analysis of the images linked by geolocated tweets. A prototype system has been built and tested on a case study based on the Tweets posted in the two days after the earthquake that occurred in Central Italy in August 2016. Results indicate that focusing on images linked by geolocated tweets represents a good criterion to identify useful information that can aid emergency response.

E2mC: Improving Emergency Management Service Practice through Social Media and Crowdsourcing Analysis in Near Real Time

“…Besides the attached image, the describing text can also be analysed to disambiguate the location of the social media post. The feasibility of geolocating images shared on Twitter posts was shown by extracting implicit geographical references from the text of the tweet and highlight a correlation existing between text features and image features [ 64 ]. Besides the challenges related to the usefulness of the extracted locations, like their precision, accuracy and credibility, the extraction phase introduces errors (false positives and false negatives), due to ambiguities which exist between location names and other proper or common names [ 65 ].…”

Exploratory Spatio-Temporal Queries in Evolving Information

“…Part of the project is focusing on extracting useful information from tweets in the case of earthquakes emergencies with adequate tools. The IMEXT [11] tool environment is a first prototype developed in the project to support Twitter crawling with specific keywords for given event types, geotagging tweets, and extracting images from tweets and from documents linked to the tweets themselves, such as other social media and traditional media.…”

“…In [13], as a case study, the earthquake in Central Italy of August 2016 has been considered, analyzing the tweets posted just after the event focusing on image extraction of potentially useful and geolocated images. Focusing on image extraction in this context, the goal is to find useful images.…”

“…In this event, using the crawling techniques illustrated in [11], about 150,000 possibly relevant tweets have been extracted from the 48 hours immediately following the event.…”

“…Concerning the images associated to the tweets, as illustrated in [11], only about 19% of the images extracted from geolocated tweets have been considered potentially useful, i.e., showing damages. As a document component we also considered videos.…”