Tri3D: Unified Interface for Multiple 3D Driving Data Sources

Tri3D is a unified interface for accessing and processing major 3D autonomous driving datasets, including KITTI, nuScenes, and the Waymo Open Dataset. Developed by CEA-LIST and last updated in November 2025, it provides a standardized framework for handling multi-modal sensor data and 3D geometry.

Use Cases

• Cross-dataset training for 3D object detection using unified labels from KITTI and nuScenes
• Benchmarking geometry-based AI models across different LiDAR sensor configurations
• Developing domain adaptation algorithms for autonomous driving using multi-source point cloud data

Strengths

• Integrates three distinct major driving datasets: KITTI, nuScenes, and Waymo Open Dataset
• Standardized geometry and coordinate system handling for 3D point clouds
• Developed by CEA-LIST, a specialized research organization in digital systems

Limitations

• Dependent on the availability and local hosting of external source datasets
• Abstraction layer may omit dataset-specific metadata not common to all three sources
• Requires manual alignment of disparate license terms from multiple original authors

Provenance

Source

CEA-LIST

Collection Method

Aggregation and unification of existing 3D driving datasets

Freshness

Last updated 2025-11-18.

Geography

Global (based on source datasets from USA, Germany, and Singapore)