Software and Datasets

CADC - Canadian Adverse Driving Conditions Dataset


Code for 3D object detection for autonomous driving.


Code to reproduce results of paper "A Bayesian Approach for Uncertainty Estimation in Deep Object Detectors".


Fast, structure-preserving depth completion on the CPU using classical morphological image processing techniques.

AscTec Pelican Quadrotor Flight Dataset

Quadrotor flight dataset created for system identification and modeling of a quadrotor for multi-step prediction.

WAVELab Software

Open source software from the former WAVELab.


Code for Monocular 3D Object Detection.


Reusable code for estimation, kinematics, optimization, vision, and more...

SIVO - Semantically Informed Visual Odometry

Integrated Bayesian semantic segmentation with ORBSLAM_2 to select better features for Visual SLAM.


Manifold geometry with fast automatic derivatives and coordinate frame semantics checking.

©2020 Toronto Robotics and AI Laboratory