Visual realism is defined as the extent to which an image appears to people as a photo rather than computer generated. Assessing visual realism is important in applications like computer graphics rendering and photo retouching. However, current realism evaluation approaches use either labor-intensive human judgments or automated algorithms largely dependent on comparing renderings to reference images. We develop a reference-free computational framework for visual realism prediction to overcome these constraints. First, we construct a benchmark dataset of 2,520 images with comprehensive human annotated attributes. From statistical modeling on this data, we identify image attributes most relevant for visual realism. We propose both empirically-based (guided by our statistical modeling of human data) and deep convolutional neural network models to predict visual realism of images. Our framework has the following advantages: (1) it creates an interpretable and concise empirical model that characterizes human perception of visual realism; (2) it links computational features to latent factors of human image perception.

Educational technological applications, such as computerized learning environments and robot tutors, are often programmed to provide social cues for the purposes of facilitating natural interaction and enhancing productive outcomes. However, there can be potential costs to social interactions that could run counter to such goals. Here, we present an experiment testing the impact of a watchful versus non-watchful robot tutor on children's language-learning effort and performance. Across two interaction sessions, children learned French and Latin rules from a robot tutor and filled in worksheets applying the rules to translate phrases. Results indicate better performance on the worksheets in the session in which the robot looked away from, as compared to the session it looked toward the child, as the child was filling in the worksheets. This was the case in particular for the more difficult worksheet items. These findings highlight the need for careful implementation of social robot behaviors to avoid counterproductive effects.

Assessing the visual realism of images is increasingly becoming an essential aspect of fields ranging from computer graphics (CG) rendering to photo manipulation. In this paper we systematically evaluate factors underlying human perception of visual realism and use that information to create an automated assessment of visual realism. We make the following unique contributions. First, we established a benchmark dataset of images with empirically determined visual realism scores. Second, we identified attributes potentially related to image realism, and used correlational techniques to determine that realism was most related to image naturalness, familiarity, aesthetics, and semantics. Third, we created an attributes-motivated, automated computational model that estimated image visual realism quantitatively. Using human assessment as a benchmark, the model was below human performance, but outperformed other state-of-the-art algorithms.

Computer-generated (CG) face images are common in video games, advertisements, and other media. CG faces vary in their degree of realism, a factor that impacts viewer reactions. Therefore, efficient control of visual realism of face images is important. Efficient control is enabled by a deep understanding of visual realism perception: the extent to which viewers judge an image as a real photograph rather than a CG image. Across two experiments, we explored the processes involved in visual realism perception of face images. In Experiment 1, participants made visual realism judgments on original face images, inverted face images, and images of faces that had the top and bottom halves misaligned. In Experiment 2, participants made visual realism judgments on original face images, scrambled faces, and images that showed different parts of faces. Our findings indicate that both holistic and piecemeal processing are involved in visual realism perception of faces, with holistic processing becoming more dominant when resolution is lower. Our results also suggest that shading information is more important than color for holistic processing, and that inversion makes visual realism judgments harder for realistic images but not for unrealistic images. Furthermore, we found that eyes are the most influential face part for visual realism, and face context is critical for evaluating realism of face parts. To the best of our knowledge, this work is a first realism-centric study attempting to bridge the human perception of visual realism on face images with general face perception tasks.

Computerized learning environments often include implementations of simulated social agents, incorporating the reasonable design assumption that learning interactions that are more social are more effective. However, recent evidence suggests that computerized social agents can in some circumstances fail to promote or even hinder learning. The current paper outlines evidence both supporting and arguing against the utility of computerized social agents for learning. We propose a framework reconciling this evidence by delineating impacts from a social agent that impinge upon different points of learning, from shallow to deep phases. Shallow social impacts when ineffectual carry little to no potential to actively impede learning, but any potential positive impacts on shallow learning phases are relatively limited in the absence of positive deep impacts. Deep social impacts, on the other hand, carry the potential to strongly drive deep learning, but when ineffectual carry the risk to impede it. The paper concludes with a proposal for future research based on this framework.