Identifying Witness Accounts from Social Media Using Imagery

Marie Truelove,Kourosh Khoshelham,Stephan Winter,Simon Mclean,Maria Vasardani

doi:10.3390/ijgi6040120

Marie Truelove, Kourosh Khoshelham + Show 3 more

Open Access

https://doi.org/10.3390/ijgi6040120

Copy DOI

Abstract

This research investigates the use of image category classification to distinguish images posted to social media that are Witness Accounts of an event. Only images depicting observations of the event, captured by micro-bloggers at the event, are considered Witness Accounts. Identifying Witness Accounts from social media is important for services such as news, marketing and emergency response. Automated image category classification is essential due to the large number of images on social media and interest in identifying witnesses in near real time. This paper begins research of this emerging problem with an established procedure, using a bag-of-words method to create a vocabulary of visual words and classifier trained to categorize the encoded images. In order to test the procedure, a set of images were collected for case study events, Australian Football League matches, from Twitter. Evaluation shows an overall accuracy of 90% and precision and recall for both classes exceeding 83%.

Highlights

Distinguishing Witness Accounts (WA) of events from social media is of significant interest for applications such as emergency services, marketing and journalism [1]
It was predicted that the number of WA OTG for the Grand Final would be similar or exceed that held on ANZAC Day, this was not the result with 54 images detected versus 97 images
The analysis demonstrated that not all images located at the Melbourne Cricket Ground (MCG) were categorized as WA OTG, and images categorized as WA OTG were posted from locations outside the venue

Summary

Introduction

Distinguishing Witness Accounts (WA) of events from social media is of significant interest for applications such as emergency services, marketing and journalism [1]. WA can provide relevant and timely information about an unfolding crisis contributing to improved situation awareness for organizations responsible for response and relief efforts [2]. They can be analyzed for feedback and sentiment related to consumer products or harvested to create event summaries [3]. With their widespread popularity, social networks are a valuable potential source for identifying witnesses [4]. A micro-blogger at a football match linking to an image of the pre-game banner (see Figure 1) is considered to have posted a WA from OTG

Objectives

Methods

Results

Discussion

Conclusion