CLOVER: Context-aware Long-term Object Viewpoint- and Environment- Invariant Representation Learning

TL;DR

This paper introduces CLOVER, a novel representation learning method for object re-identification in outdoor environments, capable of handling viewpoint and lighting variations without foreground segmentation, supported by a large diverse dataset.

Contribution

It presents CLOVER and MapCLOVER, new methods for invariant object re-identification, along with CODa Re-ID, a large dataset capturing diverse outdoor conditions.

Findings

CLOVER outperforms existing methods in re-identification accuracy.

The dataset enables robust evaluation across lighting and viewpoint changes.

CLOVER generalizes well to unseen objects and classes.

Abstract

Mobile service robots can benefit from object-level understanding of their environments, including the ability to distinguish object instances and re-identify previously seen instances. Object re-identification is challenging across different viewpoints and in scenes with significant appearance variation arising from weather or lighting changes. Existing works on object re-identification either focus on specific classes or require foreground segmentation. Further, these methods, along with object re-identification datasets, have limited consideration of challenges such as outdoor scenes and illumination changes. To address this problem, we introduce CODa Re-ID: an in-the-wild object re-identification dataset containing 1,037,814 observations of 557 objects across 8 classes under diverse lighting conditions and viewpoints. Further, we propose CLOVER, a representation learning method for object observations that can distinguish between static object instances without requiring foreground segmentation. We also introduce MapCLOVER, a method for scalably summarizing CLOVER descriptors for use in object maps and matching new observations to summarized descriptors. Our results show that CLOVER achieves superior performance in static object re-identification under varying lighting conditions and viewpoint changes and can generalize to unseen instances and classes.