CornerNet: Detecting Objects as Paired Keypoints. Law, H. & Deng, J. abstract bibtex 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 42.1% AP on MS COCO, outperforming all existing one-stage detectors.
@article{cornernet,
title = {{{CornerNet}}: {{Detecting Objects}} as {{Paired Keypoints}}},
author = {Law, Hei and Deng, Jia},
abstract = {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 42.1\% AP on MS COCO, outperforming all existing one-stage detectors.},
langid = {english},
file = {/Users/unaiaguinaco/Zotero/storage/KG7FXV4B/Law y Deng - CornerNet Detecting Objects as Paired Keypoints.pdf}
}
Downloads: 0
{"_id":"LZe9evdy6iHWYgk9y","bibbaseid":"law-deng-cornernetdetectingobjectsaspairedkeypoints","author_short":["Law, H.","Deng, J."],"bibdata":{"bibtype":"article","type":"article","title":"CornerNet: Detecting Objects as Paired Keypoints","author":[{"propositions":[],"lastnames":["Law"],"firstnames":["Hei"],"suffixes":[]},{"propositions":[],"lastnames":["Deng"],"firstnames":["Jia"],"suffixes":[]}],"abstract":"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 42.1% AP on MS COCO, outperforming all existing one-stage detectors.","langid":"english","file":"/Users/unaiaguinaco/Zotero/storage/KG7FXV4B/Law y Deng - CornerNet Detecting Objects as Paired Keypoints.pdf","bibtex":"@article{cornernet,\n title = {{{CornerNet}}: {{Detecting Objects}} as {{Paired Keypoints}}},\n author = {Law, Hei and Deng, Jia},\n abstract = {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 42.1\\% AP on MS COCO, outperforming all existing one-stage detectors.},\n langid = {english},\n file = {/Users/unaiaguinaco/Zotero/storage/KG7FXV4B/Law y Deng - CornerNet Detecting Objects as Paired Keypoints.pdf}\n}\n\n","author_short":["Law, H.","Deng, J."],"key":"cornernet-1","id":"cornernet-1","bibbaseid":"law-deng-cornernetdetectingobjectsaspairedkeypoints","role":"author","urls":{},"metadata":{"authorlinks":{}},"downloads":0,"html":""},"bibtype":"article","biburl":"https://bibbase.org/network/files/sgRwMtA3kBKedfSYM","dataSources":["8t5qpXywRa3PGo5tj"],"keywords":[],"search_terms":["cornernet","detecting","objects","paired","keypoints","law","deng"],"title":"CornerNet: Detecting Objects as Paired Keypoints","year":null}