Open-Set Object Recognition Using Mechanical Properties During Interaction

TL;DR

This paper introduces an open-set object recognition framework using mechanical properties during interaction, enabling robots to identify known objects and incrementally learn about new ones with improved clustering and recognition accuracy.

Contribution

It presents a novel clustering algorithm that leverages knowledge of known objects to improve open-set recognition and clustering performance in tactile robotics.

Findings

The framework outperforms alternative methods in object recognition accuracy.

The proposed clustering algorithm yields better clustering performance than existing methods.

Proper tuning of cluster size hyperparameters is crucial for optimal results.

Abstract

while most of the tactile robots are operated in close-set conditions, it is challenging for them to operate in open-set conditions where test objects are beyond the robots' knowledge. We proposed an open-set recognition framework using mechanical properties to recongise known objects and incrementally label novel objects. The main contribution is a clustering algorithm that exploits knowledge of known objects to estimate cluster centre and sizes, unlike a typical algorithm that randomly selects them. The framework is validated with the mechanical properties estimated from a real object during interaction. The results show that the framework could recognise objects better than alternative methods contributed by the novelty detector. Importantly, our clustering algorithm yields better clustering performance than other methods. Furthermore, the hyperparameters studies show that cluster size is important to clustering results and needed to be tuned properly.