CSCPR: Cross-Source-Context Indoor RGB-D Place Recognition

Abstract-- We present a new algorithm, Cross-Source-Context Place Recognition (CSCPR), for RGB-D indoor place recognition that integrates global retrieval and reranking into a single end-to-end model. Unlike prior approaches that primarily focus on the RGB domain, CSCPR is designed to handle the RGB-D data. We extend the Context-of-Clusters (CoCs) for handling noisy colorized point clouds and introduce two novel modules for reranking: the Self-Context Cluster (SCC) and Cross Source Context Cluster (CSCC), which enhance feature representation and match query-database pairs based on local features, respectively. Our experiments demonstrate that CSCPR significantly outperforms state-ofthe-art models on these datasets by at least 36.5% in Recall@1 at ScanNet-PR dataset and 44% in new datasets. Given a query frame, global retrieval ranks the potentially matched frames, and Place recognition plays an important role in robotics [1], our novel place-recognition reranking model reranks the [2], [3], [4], where given query frames the goal is to identify candidates to achieve better recognition accuracy. It is less computational cost.