No abstract

No abstract

No abstract

Physical adversarial attacks threaten to fool object detection systems, but reproducible research on the real-world effectiveness of physical patches and how to defend against them requires a publicly available benchmark dataset. We present APRICOT, a collection of over 1,000 annotated photographs of printed adversarial patches in public locations. The patches target several object categories for three COCO-trained detection models, and the photos represent natural variation in position, distance, lighting conditions, and viewing angle. Our analysis suggests that maintaining adversarial robustness in uncontrolled settings is highly challenging, but it is still possible to produce targeted detections under white-box and sometimes blackbox settings. We establish baselines for defending against adversarial patches through several methods, including a detector supervised with synthetic data and unsupervised methods such as kernel density estimation, Bayesian uncertainty, and reconstruction error. Our results suggest that adversarial patches can be effectively flagged, both in a high-knowledge, attack-specific scenario, and in an unsupervised setting where patches are detected as anomalies in natural images. This dataset and the described experiments provide a benchmark for future research on the effectiveness of and defenses against physical adversarial objects in the wild.

This presentation will summarize recent work on the visual perception of color appearance and object properties in optical see-through (OST) augmented reality (AR) systems. OST systems, such as Microsoft HoloLens, use a see- through display system to superimpose virtual content onto a user’s view of the real world. With careful tracking of both display and world coordinates, synthetic objects can be added to the real world, and real objects can be manipulated via synthetic overlays. Ongoing research studies how the combination of real and virtual stimuli are perceived and how users’ visual adaptation is affected; two specific examples will be explained.