CornerNet: Detecting Objects as Paired Keypoints

By Hei Law, Jia Deng. 2018

4,142 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Rethinking ImageNet Pre-Training (2018, background): “A recent work of CornerNet [22] reported 42.1 bbox AP (w/ multi-scale augmentation) using no ImageNet pre-training.”
Bridging the Gap Between Anchor-Based and Anchor-Free Detection via Adaptive Training Sample Selection (2019, methodology): “CornerNet [26] detects an object bounding box as a pair of keypoints (top-left corner and bottom-right corner) and CornerNet-Lite [27] introduces CornerNet-Saccade and CornerNet-Squeeze to improve its speed.”
Cascade R-CNN: High Quality Object Detection and Instance Segmentation (2019, methodology): “G-RMI [28], the recent Deformable R-FCN [8], RetinaNet [36], Mask R-CNN [24], RelationNet [27], DetNet [34], CornerNet [32], etc.”
Objects as Points (2019, methodology): “Hourglass The stacked Hourglass Network [30, 40] downsamples the input by 4 × , followed by two sequential hourglass modules.”
FCOS: Fully Convolutional One-Stage Object Detection (2019, background): “CornerNet requires much more complicated post-processing to group the pairs of corners belonging to the same instance.”
Exploring Plain Vision Transformer Backbones for Object Detection (2022, neutral): “anchor-free [30,14,49], and region-based [20,25,19,46] vs.”
Towards Real-World Prohibited Item Detection: A Large-Scale X-ray Benchmark (2021, background): “Recently, anchor-free approaches have at-tracted wide attention of researchers by using key points to represent the objects, including CornerNet [15], CenterNet [6], and FCOS [34].”
EfficientDet: Scalable and Efficient Object Detection (2019, background): “Recently, one-stage detectors have attracted substantial attention due to their efﬁciency and simplicity [14, 34, 36].”
NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection (2019, neutral): “0 CornerNet Hourglass [18] 512× 512 244 (Titan X) 40.”
Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection (2020, neutral): “one-stage: CornerNet [14] HG-104 200 X 3.”
Incremental Few-Shot Object Detection (2020, neutral): “It is inspired by keypoint detection methods [33, 52], taking a similar spirit to [25, 57] without the need for point grouping and postprocessing.”
Class-incremental Learning via Deep Model Consolidation (2019, methodology): “We perform all experiments using RetinaNet [30], but the proposed method is applicable to other one-stage detectors [32, 43, 54, 25] with minor modifications.”
Deep High-Resolution Representation Learning for Visual Recognition (2019, neutral): “p Regionlets [160] ResNet-101 800 1:5 39:3 59:8 - 21:7 43:7 50:9 FitnessNMS [140] ResNet-101 768 - 39:5 58:0 42:6 18:9 43:5 54:1 DetNet [86] DetNet59-FPN 800 2 40:3 62:1 43:8 23:6 42:6 50:0 CornerNet [79] Hourglass-104 511 - 40:5 56:5 43:1 19:4 42:7 53:9 M2Det [185] VGG16 800 ˘ 10 41:0 59:7 45:0 22:”
Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation (2019, background): “CornerNet [38] detects objects by predicting paired corners and group corners based on [56].”
Vision Transformers for Dense Prediction (2021, background): “Fully-convolutional networks [35, 36] are the prototypical architecture for pixel-level dense prediction tasks such as semantic segmentation [7, 25, 55], monocular depth estimation [12, 16, 32], and keypoint detection [21, 57].”
CenterNet: Keypoint Triplets for Object Detection (2019, methodology): “This paper uses CornerNet [21] as the baseline.”
YOLOv4: Optimal Speed and Accuracy of Object Detection (2020, methodology): “The detectors of this sort are CenterNet [13], CornerNet [37, 38], FCOS [78], etc.”
Libra R-CNN: Towards Balanced Learning for Object Detection (2019, background): “CornerNet [18] introduced an insight that the bounding boxes can be predicted as a pair of key points.”
Bottom-Up Object Detection by Grouping Extreme and Center Points (2019, methodology): “The standard deviation is either fixed, or set proportional to the object size [22].”