Deadpan artefacts: creative experiments with facial expression recognition in photographic portraiture
Abstract
Over the past three decades, photography has evolved from its traditional role as an indexical imprint of the world into a fluid, computational space shaped by digital transformation. Within this context, the photographic portrait, both as historical art form and a site for computer vision, offers a compelling framework for practice-led research exploring the shifting nature of photographic representation. This research explores the datafication of pose and expression through creative experiments engaging with Facial Expression Recognition (FER) technologies. A key focus is the neutral category in FER, which closely aligns with the deadpan expression, a dominant convention in fine art and documentary portraiture since the 1920s.
The collected body of practice-led research moves from using FER tools in the photographic studio to statistical analysis and visualisation of a portrait dataset. It culminates in a discursive video essay that critically synthesises the paradox of the portrait photography both as cultural form and as data. Through this research, the deadpan and the neutral have emerged as critical devices that unite the socio-technical complexities of portraiture in both historical and contemporary contexts, particularly in relation to computer vision. These recurring devices offer a distinctive framework for examining the transformation of portraiture, underscoring that expression, whether captured by a camera or interpreted by an algorithm, is never a neutral observation.
Description
Keywords
Citation
Collections
Source
Type
Book Title
Entity type
Access Statement
License Rights
DOI
Restricted until
Downloads
File
Description
Thesis Material