Computer Vision Benchmark Dataset

Computer Vision Benchmark Dataset

Research Project
Project Goals

​Taking advantage of our autonomous driving platform, we are creating a comprehensive repository of annotated sensor data that provides computer vision benchmark and training data to support advanced driving assist and autonomous driving applications.   Our autonomous vehicle platform is equipped with a variety of sensors, including cameras, laser scanners, differential GPS, radar, CAN, etc. The project goal is to collect  and annotate driving data on public roads under various driving and environment conditions to support development and evaluation of perception, planning, and control technologies.   A challenging and important aspect of this project will be development of high quality annotations using quality controlled crownd-sourcing techniques as well as semi-automated methods that take advantage of the multi-sensory information proecessing.

Research Area