The Effect of Image Presentation Rate on Person Identification

Abstract

Our ability to recognise complex images across contexts depends on our exposure to similar instances. For example, despite much natural variation, it is easier to recognise a new instance of a familiar face than an unfamiliar face. As we encounter similar images, we automatically notice structural commonalities and form a representation of how the image generally looks, even when each image is presented rapidly (i.e., several milliseconds each). However, it is not clear whether this process allows us to better identify new instances of an image compared to assessing single images for a longer duration. Across two experiments, I tested observers’ person recognition ability when presented with rapid image streams at varying rates compared to a single image. Experiment 1 compares performance between upright and inverted faces. Experiment 2 compares performance between fingerprints from the same finger and from the same person more generally. My results suggest that viewing images rapidly is better than single images when identifying faces, but not fingerprints; and that people better recognise upright compared to inverted faces, but are similar in both fingerprint conditions. I discuss the theoretical implications of these results, as well as some practical implications in security and forensic contexts. Keywords: Visual cognition, recognition, gist perception, ensemble coding, face processing, fingerprint analysis