Psychology Brown Bag: Training expertise in scene recognition

Dr. Assaf Harel, Wright State University

Training expertise in scene recognition

Abstract:

Visual analysis of complex real-world scenes is essential to a variety of professional contexts, ranging from defense and intelligence to architecture and urban planning. Expertise in recognizing information-rich yet highly variable scenes is putatively achieved through experience, yet little is currently known about how skills in scene recognition are formed and evolve during learning, and what underlying neural mechanisms support their acquisition. The present study is a first attempt at addressing these questions,
quantifying the behavioral changes associated with the acquisition of scene expertise. We assembled a rich stimulus-set consisting of high-resolution color scene images varying across five dimensions: Viewpoint (aerial/terrestrial), Naturalness (manmade/natural), and three hierarchical categorization levels: Basic-level, Subordinate, and Exemplar. For instance, the category ³deserts² contained three deserts types (Sandy, Shrub and Rocky), and each desert type contained ten individual images of specific deserts. Critically, each individual scene was presented both in an aerial and terrestrial viewpoint, to assess generalization across viewpoints. We trained 15 participants to categorize these scenes for a total of 12 hours. Each individual training regimen was comprised of six sessions; participants trained on half of the stimuli for five sessions, and in the sixth session they viewed he other half of the scenes. To assess the efficiency of training, we employed two behavioral metrics: (1) within-set learning (i.e. learning across the five sessions), and (2) generalization (i.e. transfer of learning). Learning occurred within the five sessions (evident in a monotonic decrease in reaction times and increase in accuracy), and notably, we also found transfer of learning, as performance in the sixth session was pronouncedly better than performance in the first four training sessions. Together, these results suggest that expertise in scene recognition can be trained in the lab and will form the basis for future studies on the neural substrates of scene expertise.