Fit to Measure: Reasoning about Sizes for Robust Object Recognition

摘要

Service robots can help with many of our daily tasks, especially in those cases where it is inconvenient orunsafe for us to intervene – e.g., under extreme weather conditions or when social distance needs to bemaintained. However, before we can successfully delegate complex tasks to robots, we need to enhancetheir ability to make sense of dynamic, real-world environments. In this context, the first prerequisite toimproving the Visual Intelligence of a robot is building robust and reliable object recognition systems.While object recognition solutions are traditionally based on Machine Learning, augmenting them withknowledge-based reasoners has been shown to improve their performance. In particular, based on ourprior work on identifying the epistemic requirements of Visual Intelligence, we hypothesise that knowl-edge of the typical size of objects can significantly improve the accuracy of an object recognition system.To verify this hypothesis, in this paper we present an approach to integrating knowledge about objectsizes in a ML-based architecture. Our experiments in a real-world robotic scenario show that this hybridapproach ensures a significant performance increase over state-of-the-art Machine Learning methods.