Max Robotics — Global Service Robot Directory + US Certification

About

John Skinner is a researcher whose work sits at the intersection of robotic vision, probabilistic object detection, and computer vision evaluation methodologies. His research addresses a fundamental challenge in robotics: the difficulty of replicating experimental conditions when cameras operate in dynamic environments with shifting lighting and moving objects. His 2016 paper on high-fidelity simulation for evaluating robotic vision performance (20 citations) proposed simulation as a rigorous alternative for benchmarking, offering reproducibility that real-world testing cannot guarantee. Skinner is perhaps best known for pioneering the concept of Probabilistic Object Detection — the task of detecting objects while accurately quantifying spatial and semantic uncertainties. His development of the Probability-based Detection Quality (PDQ) metric, introduced in 2018 and expanded in a 2020 paper (collectively gathering nearly 30 citations), gave the research community its first principled framework for evaluating such uncertainty-aware detections. This work culminated in the ACRV Robotic Vision Challenge, which he helped organize to push the field toward more honest, uncertainty-conscious detection systems. His 2022 reflective paper examining what robotics research can learn from computer vision underscores his broader commitment to bridging these disciplines and improving how robotic systems are developed, evaluated, and compared.