TY - GEN
T1 - Radatron
T2 - 17th European Conference on Computer Vision, ECCV 2022
AU - Madani, Sohrab
AU - Guan, Jayden
AU - Ahmed, Waleed
AU - Gupta, Saurabh
AU - Hassanieh, Haitham
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - Millimeter wave (mmWave) radars are becoming a more popular sensing modality in self-driving cars due to their favorable characteristics in adverse weather. Yet, they currently lack sufficient spatial resolution for semantic scene understanding. In this paper, we present Radatron, a system capable of accurate object detection using mmWave radar as a stand-alone sensor. To enable Radatron, we introduce a first-of-its-kind, high-resolution automotive radar dataset collected with a cascaded MIMO (Multiple Input Multiple Output) radar. Our radar achieves 5 cm range resolution and 1.2 ∘ angular resolution, 10 × finer than other publicly available datasets. We also develop a novel hybrid radar processing and deep learning approach to achieve high vehicle detection accuracy. We train and extensively evaluate Radatron to show it achieves 92.6 % AP 50 and 56.3 % AP 75 accuracy in 2D bounding box detection, an 8 % and 15.9 % improvement over prior art respectively. Code and dataset is available on https://jguan.page/Radatron/.
AB - Millimeter wave (mmWave) radars are becoming a more popular sensing modality in self-driving cars due to their favorable characteristics in adverse weather. Yet, they currently lack sufficient spatial resolution for semantic scene understanding. In this paper, we present Radatron, a system capable of accurate object detection using mmWave radar as a stand-alone sensor. To enable Radatron, we introduce a first-of-its-kind, high-resolution automotive radar dataset collected with a cascaded MIMO (Multiple Input Multiple Output) radar. Our radar achieves 5 cm range resolution and 1.2 ∘ angular resolution, 10 × finer than other publicly available datasets. We also develop a novel hybrid radar processing and deep learning approach to achieve high vehicle detection accuracy. We train and extensively evaluate Radatron to show it achieves 92.6 % AP 50 and 56.3 % AP 75 accuracy in 2D bounding box detection, an 8 % and 15.9 % improvement over prior art respectively. Code and dataset is available on https://jguan.page/Radatron/.
UR - http://www.scopus.com/inward/record.url?scp=85142758937&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85142758937&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-19842-7_10
DO - 10.1007/978-3-031-19842-7_10
M3 - Conference contribution
AN - SCOPUS:85142758937
SN - 9783031198410
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 160
EP - 178
BT - Computer Vision – ECCV 2022 - 17th European Conference, Proceedings
A2 - Avidan, Shai
A2 - Brostow, Gabriel
A2 - Cissé, Moustapha
A2 - Farinella, Giovanni Maria
A2 - Hassner, Tal
PB - Springer
Y2 - 23 October 2022 through 27 October 2022
ER -