Large-scale geo-facial image analysis

Abstract

While face analysis from images is a well-studied area, little work has explored the dependence of facial appearance on the geographic location from which the image was captured. To fill this gap, we constructed GeoFaces, a large dataset of geotagged face images, and used it to examine the geo-dependence of facial features and attributes, such as ethnicity, gender, or the presence of facial hair. Our analysis illuminates the relationship between raw facial appearance, facial attributes, and geographic location, both globally and in selected major urban areas. Some of our experiments, and the resulting visualizations, confirm prior expectations, such as the predominance of ethnically Asian faces in Asia, while others highlight novel information that can be obtained with this type of analysis, such as the major city with the highest percentage of people with a mustache.

Walk and Learn: Facial Attribute Representation Learning from Egocentric Video and Contextual Data

“…Other related research includes discovering architectural elements and recognizing city attributes from large geo-tagged image repositories [8,59] and using location context to improve image classification [46]. More closely related to our work, Islam et al [16,17] investigated the geo-dependence of facial features and attributes; however they used off-the-shelf facial attribute classifiers for this analysis, whereas the goal of our work is to build feature representations so as to improve the accuracy of facial attribute classifiers.…”

Assemblage of the vertical: commercial drones and algorithmic life

“…Facial detection, tracking and recognition is fast becoming a significant factor in security discourses. One recent study for example sought to identify the most typical types of faces in different geographical areas (Islam et al, 2015). It would be important to further study how drone-enabled facial recognition could then identify individuals who were "out of place" in certain situations.…”

StreetStyle: Exploring world-wide clothing styles from millions of photos

“…Other work has used large image collections to study people, including explorations of how people move through cities [Crandall et al 2009], the relationship between facial appearance and geolocation/time [Islam et al 2015;Salem et al 2016], analysis of expressions and styles over a century in high school yearbook photos [Ginosar et al 2015], and tools for discovering visual patterns that distinguish two populations [Matzen and Snavely 2015]. Finally, Gebru et al study demographics across the US by detecting and analyzing cars (along with their makes and models) in Street View imagery [2017].…”