Learning Global Representation from Queries for Vectorized HD Map Construction

2025-10-08

AI Summary

Existing DETR-based online vectorized high-definition (HD) map construction methods rely on independent, learnable object queries, limiting modeling to local instances and hindering the representation of global geometric semantics—such as road topology. To address this, we propose a Global Representation Learning (GRL) and Guidance (GRG) framework that explicitly models inter-query global relationships within the DETR architecture: (i) query distributions are aligned via holistic segmentation supervision, and (ii) global contextual information is injected into each query, enabling joint optimization of local detection and global structural modeling. Evaluated on nuScenes and Argoverse2, our method substantially outperforms mainstream baselines, achieving significant mAP improvements. These results empirically validate the critical role of explicit global representation modeling in HD map vectorization.

Technology Category

• Computer Vision: Representation Learning for Vision
• Machine Learning: Representation Learning
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

The online construction of vectorized high-definition (HD) maps is a cornerstone of modern autonomous driving systems. State-of-the-art approaches, particularly those based on the DETR framework, formulate this as an instance detection problem. However, their reliance on independent, learnable object queries results in a predominantly local query perspective, neglecting the inherent global representation within HD maps. In this work, we propose extbf{MapGR} ( extbf{G}lobal extbf{R}epresentation learning for HD extbf{Map} construction), an architecture designed to learn and utilize a global representations from queries. Our method introduces two synergistic modules: a Global Representation Learning (GRL) module, which encourages the distribution of all queries to better align with the global map through a carefully designed holistic segmentation task, and a Global Representation Guidance (GRG) module, which endows each individual query with explicit, global-level contextual information to facilitate its optimization. Evaluations on the nuScenes and Argoverse2 datasets validate the efficacy of our approach, demonstrating substantial improvements in mean Average Precision (mAP) compared to leading baselines.

Problem

• Learning global representations from queries for HD map construction
• Addressing local perspective limitations in autonomous driving systems
• Enhancing query optimization with global contextual information

Innovation

• Learning global representation from queries
• Introducing Global Representation Learning module
• Adding Global Representation Guidance module