Matches in SemOpenAlex for { <https://semopenalex.org/work/W3215623930> ?p ?o ?g. }
- W3215623930 abstract "The target of 2D human pose estimation is to locate the keypoints of body parts from 2D images. State-of-the-art methods for pose estimation usually construct pixel-wise heatmaps from keypoints as labels for learning neural networks, which are usually initialized randomly or using classification models on large dataset, such as ImageNet, for their backbones. According to statistical data, there are strong positional priors for human keypoints, which are highly dependent on their relationship between image patches. To learn positional priors for pretraining pose estimators, we propose Heatmap-Style Jigsaw Puzzles (HSJP) problem as self-supervised pretext task, whose target is to predict the location of each patch from an image composed of shuffled patches. During pretraining, we only use person images in MS-COCO, rather than introducing extra large dataset like ImageNet. A heatmap-style label for patch location is designed and our learning process is in a non-contrastive way. The weights learned by HSJP pretext task are utilised as backbones of 2D human pose estimators, which are then finetuned on MS-COCO human keypoints dataset. With two popular and strong 2D human pose estimators, HRNet and SimpleBaseline, we evaluate mAP score on both MS-COCO validation and test-dev datasets. Our experiments show that downstream pose estimators with our self-supervised pretraining obtain much better performance than those trained from scratch, and are comparable to those using ImageNet classification models as their initial backbones." @default.
- W3215623930 created "2021-12-06" @default.
- W3215623930 creator A5005122591 @default.
- W3215623930 creator A5015146540 @default.
- W3215623930 creator A5016038041 @default.
- W3215623930 creator A5019499419 @default.
- W3215623930 creator A5019797064 @default.
- W3215623930 creator A5033273449 @default.
- W3215623930 creator A5046293929 @default.
- W3215623930 creator A5055332125 @default.
- W3215623930 creator A5087241722 @default.
- W3215623930 date "2021-11-23" @default.
- W3215623930 modified "2023-09-23" @default.
- W3215623930 title "Learning Positional Priors for Pretraining 2D Pose Estimators" @default.
- W3215623930 cites W1982931780 @default.
- W3215623930 cites W2053923031 @default.
- W3215623930 cites W2108598243 @default.
- W3215623930 cites W2113325037 @default.
- W3215623930 cites W2128745377 @default.
- W3215623930 cites W219040644 @default.
- W3215623930 cites W2207258529 @default.
- W3215623930 cites W2307770531 @default.
- W3215623930 cites W2308529009 @default.
- W3215623930 cites W2321533354 @default.
- W3215623930 cites W2326925005 @default.
- W3215623930 cites W2554082199 @default.
- W3215623930 cites W2559085405 @default.
- W3215623930 cites W2892032594 @default.
- W3215623930 cites W2916798096 @default.
- W3215623930 cites W2944341987 @default.
- W3215623930 cites W2958360136 @default.
- W3215623930 cites W2962820842 @default.
- W3215623930 cites W2963076818 @default.
- W3215623930 cites W2963402313 @default.
- W3215623930 cites W2963426332 @default.
- W3215623930 cites W2963438548 @default.
- W3215623930 cites W2964221239 @default.
- W3215623930 cites W2997471409 @default.
- W3215623930 cites W3034742259 @default.
- W3215623930 cites W3034750257 @default.
- W3215623930 cites W3034930876 @default.
- W3215623930 cites W3035004948 @default.
- W3215623930 cites W3035106315 @default.
- W3215623930 cites W3035164673 @default.
- W3215623930 cites W639708223 @default.
- W3215623930 doi "https://doi.org/10.1145/3475723.3484252" @default.
- W3215623930 hasPublicationYear "2021" @default.
- W3215623930 type Work @default.
- W3215623930 sameAs 3215623930 @default.
- W3215623930 citedByCount "0" @default.
- W3215623930 crossrefType "proceedings-article" @default.
- W3215623930 hasAuthorship W3215623930A5005122591 @default.
- W3215623930 hasAuthorship W3215623930A5015146540 @default.
- W3215623930 hasAuthorship W3215623930A5016038041 @default.
- W3215623930 hasAuthorship W3215623930A5019499419 @default.
- W3215623930 hasAuthorship W3215623930A5019797064 @default.
- W3215623930 hasAuthorship W3215623930A5033273449 @default.
- W3215623930 hasAuthorship W3215623930A5046293929 @default.
- W3215623930 hasAuthorship W3215623930A5055332125 @default.
- W3215623930 hasAuthorship W3215623930A5087241722 @default.
- W3215623930 hasConcept C105795698 @default.
- W3215623930 hasConcept C107673813 @default.
- W3215623930 hasConcept C115961682 @default.
- W3215623930 hasConcept C119857082 @default.
- W3215623930 hasConcept C153180895 @default.
- W3215623930 hasConcept C154945302 @default.
- W3215623930 hasConcept C162324750 @default.
- W3215623930 hasConcept C177769412 @default.
- W3215623930 hasConcept C185429906 @default.
- W3215623930 hasConcept C187736073 @default.
- W3215623930 hasConcept C2780451532 @default.
- W3215623930 hasConcept C31972630 @default.
- W3215623930 hasConcept C33923547 @default.
- W3215623930 hasConcept C41008148 @default.
- W3215623930 hasConcept C52102323 @default.
- W3215623930 hasConceptScore W3215623930C105795698 @default.
- W3215623930 hasConceptScore W3215623930C107673813 @default.
- W3215623930 hasConceptScore W3215623930C115961682 @default.
- W3215623930 hasConceptScore W3215623930C119857082 @default.
- W3215623930 hasConceptScore W3215623930C153180895 @default.
- W3215623930 hasConceptScore W3215623930C154945302 @default.
- W3215623930 hasConceptScore W3215623930C162324750 @default.
- W3215623930 hasConceptScore W3215623930C177769412 @default.
- W3215623930 hasConceptScore W3215623930C185429906 @default.
- W3215623930 hasConceptScore W3215623930C187736073 @default.
- W3215623930 hasConceptScore W3215623930C2780451532 @default.
- W3215623930 hasConceptScore W3215623930C31972630 @default.
- W3215623930 hasConceptScore W3215623930C33923547 @default.
- W3215623930 hasConceptScore W3215623930C41008148 @default.
- W3215623930 hasConceptScore W3215623930C52102323 @default.
- W3215623930 hasLocation W32156239301 @default.
- W3215623930 hasOpenAccess W3215623930 @default.
- W3215623930 hasPrimaryLocation W32156239301 @default.
- W3215623930 hasRelatedWork W1968716783 @default.
- W3215623930 hasRelatedWork W2119493293 @default.
- W3215623930 hasRelatedWork W2286009621 @default.
- W3215623930 hasRelatedWork W2736638679 @default.
- W3215623930 hasRelatedWork W2774989038 @default.
- W3215623930 hasRelatedWork W3108980762 @default.
- W3215623930 hasRelatedWork W4302043214 @default.