DeepPerception > Objectives

Project Objectives

The project main objective is study and development of holistic deep learning and computer vision based 3D perception solutions integrating geometry, semantics and motion from multiple sensors, making the autonomous mobile system capable of accurately parsing, reconstructing and understanding its surrounding environment. We will study three solutions, two of them being independent, one based on LiDAR, another on cameras, and the third one based on low-level fusion between the two sensors.

The estimated results of this project are:
-LiDAR-based 3D point cloud perception solution.
-Camera-based 3D point cloud perception solution.
-Multi-sensor based solutions for enhanced 3D perception.
-The solutions will be deployed and compared on an experimental vehicle.
-The results will be published in at list 3 Q1 journals and 3 high-level conferences.