Development and validation of an AI-generated real-world object stimuli set

The availability of real-world object stimuli that meet researchers’ requirements is an ongoing challenge in visual cognition research. While numerous manually curated object stimuli sets exist, stimuli features such as size, colour and orientation tend to vary widely within a given set, and may not be suitable for studies with specific requirements regarding these parameters. However, recent advances in artificial intelligence (AI) can facilitate the generation of custom-made, highly realistic stimuli. Building on these developments, the present aim was to share a set of 200 AI-generated everyday object images for use in research. The objects were oriented as though ‘placed’ on a flat surface, such that they could be naturally embedded in virtual scenes. Moreover, they were created in greyscale and suitable for rendering in different colours. Here, we report the method used to efficiently generate the stimuli, as well as the results from a validation study in which we assessed the nameability, perceived realism and familiarity of the stimuli in a sample of 45 younger (18-35) and 45 older (65-85) adults. As anticipated, the majority of the stimuli were rated highly across all three measures, and no significant age differences were observed. The results thus validated most of the stimuli for future research. The stimuli, each in seven colours, and the corresponding validation scores are openly available for future use. To our knowledge, the current work is the first to develop an openly available AI-generated real-world object stimuli set with corresponding validation measures of nameability, realism and familiarity.