ArxNet Model and Data: Building Social Networks from Image Archives

TL;DR

This paper introduces ArxNet, a model that extracts and analyzes social connections from large collections of images using face detection and social network analysis, demonstrated on Emmy's Award images.

Contribution

The paper presents a novel approach combining face detection and social network analysis to build social networks from image archives.

Findings

Successfully identified social connections in Emmy's Award images.

Demonstrated effectiveness of the method on a large public image dataset.

Provides a new tool for social network analysis using visual data.

Abstract

A corresponding explosion in digital images has accompanied the rapid adoption of mobile technology around the world. People and their activities are routinely captured in digital image and video files. By their very nature, these images and videos often portray social and professional connections. Individuals in the same picture are often connected in some meaningful way. Our research seeks to identify and model social connections found in images using modern face detection technology and social network analysis. The proposed methods are then demonstrated on the public image repository associated with the 2022 Emmy's Award Presentation.