Structuring Quantitative Image Analysis with Metric Depth

• In Event: Automated Annotation of Text, Image, and Video Data

Abstract

When photographers---and other editors of image material---produce an image, they make a statement about what matters by situating some objects in the foreground and others in the background. While this prominence of objects is a key analytical category to qualitative scholars, recent quantitative approaches to automated image analysis have not yet made this important distinction but treat all areas of an image similarly. We rely on the pixels' image depth to detect such object prominence, which allows us to combine the best of the two worlds: The conceptual precision of distinguishing between foreground and background from qualitative analyses with the scalability of quantitative approaches. Joining those two aspects makes quantitative approaches conceptually more meaningful. We showcase our approach in two applications. Adding image depth increases the analytical leverage of the framework proposed in Torres (forthcoming). In a second application, we also illustrate how to use image depth to identify visual key messages in images of news articles.

Authors

• Christian Arnold, Cardiff University

• Andreas Küpfer, Technical University of Darmstadt