Objects as Points

16 Apr 2019  ·  Xingyi Zhou, Dequan Wang, Philipp Krähenbühl ·

Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and requires additional post-processing. In this paper, we take a different approach. We model an object as a single point --- the center point of its bounding box. Our detector uses keypoint estimation to find center points and regresses to all other object properties, such as size, 3D location, orientation, and even pose. Our center point based approach, CenterNet, is end-to-end differentiable, simpler, faster, and more accurate than corresponding bounding box based detectors. CenterNet achieves the best speed-accuracy trade-off on the MS COCO dataset, with 28.1% AP at 142 FPS, 37.4% AP at 52 FPS, and 45.1% AP with multi-scale testing at 1.4 FPS. We use the same approach to estimate 3D bounding box in the KITTI benchmark and human pose on the COCO keypoint dataset. Our method performs competitively with sophisticated multi-stage methods and runs in real-time.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Object Detection COCO test-dev CenterNet-DLA (DLA-34, multi-scale) box mAP 41.6 # 176
APS 21.5 # 129
APM 43.9 # 118
APL 56.0 # 91
Hardware Burden 26G # 1
Object Detection PASCAL VOC 2007 CenterNet(DLA34, Flip, 512x512) MAP 80.7% # 11
One-stage Anchor-free Oriented Object Detection SKU110K-R CenterNet AP 54.7 # 4
AP@75 61.1 # 4
Object Detection UA-DETRAC CenterNet mAP 83.48 # 4

Methods


No methods listed for this paper. Add relevant methods here