CornerNet: Detecting Objects as Paired Keypoints
This addresses the problem of simplifying and improving object detection for computer vision applications, though it is incremental in its method innovation.
CornerNet tackles object detection by representing bounding boxes as paired keypoints (top-left and bottom-right corners), eliminating the need for anchor boxes, and achieves a 42.2% AP on MS COCO, outperforming all existing one-stage detectors.
We propose CornerNet, a new approach to object detection where we detect an object bounding box as a pair of keypoints, the top-left corner and the bottom-right corner, using a single convolution neural network. By detecting objects as paired keypoints, we eliminate the need for designing a set of anchor boxes commonly used in prior single-stage detectors. In addition to our novel formulation, we introduce corner pooling, a new type of pooling layer that helps the network better localize corners. Experiments show that CornerNet achieves a 42.2% AP on MS COCO, outperforming all existing one-stage detectors.