An Empirical Study of Photo Naming Conventions across Popular Social Media Platforms
Anh-Duy Tran, Dinh-Hai Le, Thien-Phu Dao, Minh–Triet Tran, Duc‐Tien Dang‐Nguyen · 2024
Metadata analysis is a crucial aspect of image forensics, but what happens when all information about an image, including its metadata, is lost? In such situations, it is important to explore alternative methods for image analysis. Our research has discovered that the filenames of uploaded images on social media platforms can provide valuable information. Building upon this observation, our study investigates the naming conventions and image source tracing of three popular online platforms: Facebook, Twitter, and Flickr. In addition, we identify the origin of uploaded images and extract other valuable information from the filenames. Finally, we construct a dataset consisting of 497,368 image filenames from Facebook, 13,703 image filenames with metadata from Twitter, and 1.6 million pairs of Flickr IDs with their respective posted dates.