RRPN: Radar Region Proposal Network for Object Detection in Autonomous Vehicles

1 May 2019  ·  Ramin Nabati, Hairong Qi ·

Region proposal algorithms play an important role in most state-of-the-art two-stage object detection networks by hypothesizing object locations in the image. Nonetheless, region proposal algorithms are known to be the bottleneck in most two-stage object detection networks, increasing the processing time for each image and resulting in slow networks not suitable for real-time applications such as autonomous driving vehicles. In this paper we introduce RRPN, a Radar-based real-time region proposal algorithm for object detection in autonomous driving vehicles. RRPN generates object proposals by mapping Radar detections to the image coordinate system and generating pre-defined anchor boxes for each mapped Radar detection point. These anchor boxes are then transformed and scaled based on the object's distance from the vehicle, to provide more accurate proposals for the detected objects. We evaluate our method on the newly released NuScenes dataset [1] using the Fast R-CNN object detection network [2]. Compared to the Selective Search object proposal algorithm [3], our model operates more than 100x faster while at the same time achieves higher detection precision and recall. Code has been made publicly available at https://github.com/mrnabati/RRPN .

PDF Abstract


Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
3D Object Detection nuScenes-F RRPN + R101 - F AP 43 # 1
AP50 64.9 # 1
AP75 48.5 # 1
AR 48.6 # 1
ARI 58.2 # 1
ARm 41.2 # 1
ARs 4 # 1
3D Object Detection nuScenes-FB RRPN + R101 - FB AP 35.5 # 1
AP50 59 # 1
AP75 37 # 1
AR 42.1 # 1
ARI 51.4 # 1
ARm 39.1 # 1
ARs 21.1 # 1