Prof. Kurt Keutzer’s presentation at the 2018 CVPR Workshop on Autonomous Driving
11 Challenges for Computer Vision from Autonomous Vehicles:
Solving the perception piece of the autonomous driving puzzle
1.Rethinking accuracy
2.Getting enough data
3.Generating useful data through simulation
4.Formally directed simulation
5.Domain adaption
6.Accelerating training to cope with all this new data New Deep Learning algorithms:
7.Handlinbg high resolution
8. Getting more out of video
9. Utilizing sensor fusion (LiDAR, RADAR etc)
10. Designing power and energy efficient nets
11. More efficient computation through NN Accelerators
GIPHY App Key not set. Please check settings