Kitting in the Wild through Online Domain Adaptation

TL;DR

This paper introduces a new dataset for robotic kitting under varied conditions and proposes an online adaptation method that improves visual recognition robustness without requiring target domain data during training.

Contribution

It presents a novel dataset for evaluating robustness of visual recognition in robotic kitting and an online adaptation algorithm based on batch-normalization that operates without target domain data.

Findings

The dataset enables testing of domain shift robustness in robotic vision.

The online adaptation method improves recognition performance under changing conditions.

The algorithm narrows the gap between standard and offline adapted models.

Abstract

Technological developments call for increasing perception and action capabilities of robots. Among other skills, vision systems that can adapt to any possible change in the working conditions are needed. Since these conditions are unpredictable, we need benchmarks which allow to assess the generalization and robustness capabilities of our visual recognition algorithms. In this work we focus on robotic kitting in unconstrained scenarios. As a first contribution, we present a new visual dataset for the kitting task. Differently from standard object recognition datasets, we provide images of the same objects acquired under various conditions where camera, illumination and background are changed. This novel dataset allows for testing the robustness of robot visual recognition algorithms to a series of different domain shifts both in isolation and unified. Our second contribution is a novel online adaptation algorithm for deep models, based on batch-normalization layers, which allows to continuously adapt a model to the current working conditions. Differently from standard domain adaptation algorithms, it does not require any image from the target domain at training time. We benchmark the performance of the algorithm on the proposed dataset, showing its capability to fill the gap between the performances of a standard architecture and its counterpart adapted offline to the given target domain.