CornerNet-Lite: Efficient Keypoint Based Object Detection

18 Apr 2019  ·  Hei Law, Yun Teng, Olga Russakovsky, Jia Deng ·

Keypoint-based methods are a relatively new paradigm in object detection, eliminating the need for anchor boxes and offering a simplified detection framework. Keypoint-based CornerNet achieves state of the art accuracy among single-stage detectors. However, this accuracy comes at high processing cost. In this work, we tackle the problem of efficient keypoint-based object detection and introduce CornerNet-Lite. CornerNet-Lite is a combination of two efficient variants of CornerNet: CornerNet-Saccade, which uses an attention mechanism to eliminate the need for exhaustively processing all pixels of the image, and CornerNet-Squeeze, which introduces a new compact backbone architecture. Together these two variants address the two critical use cases in efficient object detection: improving efficiency without sacrificing accuracy, and improving accuracy at real-time efficiency. CornerNet-Saccade is suitable for offline processing, improving the efficiency of CornerNet by 6.0x and the AP by 1.0% on COCO. CornerNet-Squeeze is suitable for real-time detection, improving both the efficiency and accuracy of the popular real-time detector YOLOv3 (34.4% AP at 30ms for CornerNet-Squeeze compared to 33.0% AP at 39ms for YOLOv3 on COCO). Together these contributions for the first time reveal the potential of keypoint-based detection to be useful for applications requiring processing efficiency.

PDF Abstract

Datasets


Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Object Detection COCO minival CornerNet-Saccade (Hourglass-54) box AP 42.6 # 117
APS 25.5 # 45
APM 44.3 # 55
APL 58.4 # 40
Object Detection COCO minival CornerNet-Saccade (Hourglass-104) box AP 41.4 # 130
APS 23.8 # 56
APM 43.5 # 62
APL 57.1 # 48
Object Detection COCO test-dev CornerNet-Saccade (Hourglass-104, multi-scale) box AP 43.2 # 150
APS 24.4 # 104
APM 44.6 # 116
APL 57.3 # 79
Hardware Burden None # 1
Operations per network pass None # 1
Object Detection COCO test-dev CornerNet-Squeeze box AP 34.4 # 218
Hardware Burden None # 1
Operations per network pass None # 1

Methods