What Makes Photo Cultures Different?
Co-authors: Miriam Redi (Bell Labs), Damon Crockett (UCSD), and Simon Osindero (Flickr).
ACM Multimedia 2016, October 2016.
Summary:
In sociology and media history, the term “culture” is used to characterize behaviors, beliefs or artifacts of a group of individuals in a particular time period and location(s). For example, rather than thinking of “photography” as a single phenomenon, it is more precise to consider it as a collection of many different “photo cultures”, each with its set of distinct aesthetic rules and defining mechanisms. Examples of photo cultures include the “New Vision” European photography in late 1920s, the socially conscious photography practiced in New York in the 1930s, or the snapshot-style fashion photography in the 1990s.
How can this perspective can inform studies of contemporary user-generated content? Just as photography did during its first 160 years, online photo sharing platforms such as Instagram includes many different photo cultures. However, previous publications analyzing Instagram images often approaches this medium as a single global photo monoculture. This research tends to use large samples drawn from Instagram as a whole, without considering possible differences in how Instagram is used by people with different backgrounds or geographic areas. Our hypothesis is that different users employ this medium in different ways, making Instagram a collection of photo cultures sharing pictures with different subjects and stylistic attributes. Although people who spend significant time on Instagram user may be aware of this, so far no quantitative analysis of such multiple photo cultures has been carried out.
The 2013 project Phototrails compared 2.3M Instagram images from 13 cities but only along a few visual dimensions. The 2014 project Selfiecity went further by using computer vision to compare many characteristics of selfies shared in five megacities. These earlier projects motivated the present study.
Our paper for the first time compares content, photo techniques and visual styles of Instagram images along geographic dimension. Using deep learning, we detect 1000 types of content in the dataset of 100,000 images. We also extract many features that characterizes images visual style, photo techniques and aesthetic properties. We then propose and test a few different methods for comparing image samples shared in five cities using content and visual features.
The first method uses a custom visualization technique and clustering. The second method uses supervised learning multi-class classification to quantify the differences between cities’ image samples. The third method compares the samples along stereotypical-unique dimension.