🇦🇺 WildCross: Benchmark for Visual/Cross-Modal Place Recognition

Location: Australia

WildCross is a large-scale benchmark for Visual/Cross-Modal place recognition and metric depth estimation in natural environments, featuring over 476K sequential RGB frames.

A new large-scale benchmark, WildCross, has been released for Visual/Cross-Modal place recognition and metric depth estimation in natural environments.

Key features of WildCross include:

• Over 476K sequential RGB frames with semi-dense metric depth and surface normal frames
• Aligned with accurate 6DoF poses and synchronized dense LiDAR submaps
• Supports Visual Place Recognition (VPR), Cross-Modal Place Recognition (CMPR), and metric depth estimation

Explore the project further at https://csiro-robotics.github.io/WildCross/, and access the paper on arXiv.

Tags: WildCross, Visual Place Recognition, Cross-Modal Place Recognition, metric depth estimation, robotic perception, natural environments, benchmark, Australia