Matches in SemOpenAlex for { <https://semopenalex.org/work/W4200126603> ?p ?o ?g. }
- W4200126603 endingPage "974" @default.
- W4200126603 startingPage "960" @default.
- W4200126603 abstract "Human pose estimation (HPE) is crucial for computer vision (CV). Moreover, it’s a vital step for computers to understand human actions and behaviours. However, the huge number of parameters and calculations in the HPE model have brought big challenges to deploy to resource-constrained mobile devices. Aiming to overcome the challenge, we propose a sparse pruning method (SPM) for the HPE model. First, L1 regularisation is added in the training phase of the original model, and network parameters of the convolution layers (CLs) and batch normalisation layers (BNLs) are sparsely trained to obtain a network structure with sparse weights. We then combine the sparse weights of filters with the scaling parameters of the BNLs to determine their importance. Finally, the structured pruning method is used to prune the sparse filters and corresponding channels. SPM can reduce the number of model parameters and calculations without affecting precision. Promising results indicate that SPM outperforms other advanced pruning methods." @default.
- W4200126603 created "2021-12-31" @default.
- W4200126603 creator A5019503356 @default.
- W4200126603 creator A5030581708 @default.
- W4200126603 creator A5061084605 @default.
- W4200126603 creator A5068285118 @default.
- W4200126603 creator A5071183299 @default.
- W4200126603 creator A5081503976 @default.
- W4200126603 date "2021-12-13" @default.
- W4200126603 modified "2023-10-14" @default.
- W4200126603 title "An efficient sparse pruning method for human pose estimation" @default.
- W4200126603 cites W1972247907 @default.
- W4200126603 cites W1988620791 @default.
- W4200126603 cites W1994529670 @default.
- W4200126603 cites W2036966696 @default.
- W4200126603 cites W2080873731 @default.
- W4200126603 cites W2097151019 @default.
- W4200126603 cites W2113325037 @default.
- W4200126603 cites W2194775991 @default.
- W4200126603 cites W2307770531 @default.
- W4200126603 cites W2460144244 @default.
- W4200126603 cites W2808168148 @default.
- W4200126603 cites W2892096486 @default.
- W4200126603 cites W2916798096 @default.
- W4200126603 cites W2949948129 @default.
- W4200126603 cites W2962730651 @default.
- W4200126603 cites W2962851801 @default.
- W4200126603 cites W2963363373 @default.
- W4200126603 cites W2964304707 @default.
- W4200126603 cites W2965262829 @default.
- W4200126603 cites W2977140025 @default.
- W4200126603 cites W3034513523 @default.
- W4200126603 cites W3035377608 @default.
- W4200126603 cites W3038360500 @default.
- W4200126603 cites W3084016763 @default.
- W4200126603 cites W3093480222 @default.
- W4200126603 cites W3102288316 @default.
- W4200126603 cites W3111255625 @default.
- W4200126603 cites W3115462576 @default.
- W4200126603 cites W3122431931 @default.
- W4200126603 cites W3130577181 @default.
- W4200126603 cites W3162040864 @default.
- W4200126603 cites W3171048170 @default.
- W4200126603 cites W3192863538 @default.
- W4200126603 doi "https://doi.org/10.1080/09540091.2021.2012423" @default.
- W4200126603 hasPublicationYear "2021" @default.
- W4200126603 type Work @default.
- W4200126603 citedByCount "9" @default.
- W4200126603 countsByYear W42001266032022 @default.
- W4200126603 countsByYear W42001266032023 @default.
- W4200126603 crossrefType "journal-article" @default.
- W4200126603 hasAuthorship W4200126603A5019503356 @default.
- W4200126603 hasAuthorship W4200126603A5030581708 @default.
- W4200126603 hasAuthorship W4200126603A5061084605 @default.
- W4200126603 hasAuthorship W4200126603A5068285118 @default.
- W4200126603 hasAuthorship W4200126603A5071183299 @default.
- W4200126603 hasAuthorship W4200126603A5081503976 @default.
- W4200126603 hasBestOaLocation W42001266031 @default.
- W4200126603 hasConcept C108010975 @default.
- W4200126603 hasConcept C11413529 @default.
- W4200126603 hasConcept C119767625 @default.
- W4200126603 hasConcept C119857082 @default.
- W4200126603 hasConcept C153180895 @default.
- W4200126603 hasConcept C154945302 @default.
- W4200126603 hasConcept C190729725 @default.
- W4200126603 hasConcept C41008148 @default.
- W4200126603 hasConcept C45347329 @default.
- W4200126603 hasConcept C50644808 @default.
- W4200126603 hasConcept C6557445 @default.
- W4200126603 hasConcept C71924100 @default.
- W4200126603 hasConcept C86803240 @default.
- W4200126603 hasConceptScore W4200126603C108010975 @default.
- W4200126603 hasConceptScore W4200126603C11413529 @default.
- W4200126603 hasConceptScore W4200126603C119767625 @default.
- W4200126603 hasConceptScore W4200126603C119857082 @default.
- W4200126603 hasConceptScore W4200126603C153180895 @default.
- W4200126603 hasConceptScore W4200126603C154945302 @default.
- W4200126603 hasConceptScore W4200126603C190729725 @default.
- W4200126603 hasConceptScore W4200126603C41008148 @default.
- W4200126603 hasConceptScore W4200126603C45347329 @default.
- W4200126603 hasConceptScore W4200126603C50644808 @default.
- W4200126603 hasConceptScore W4200126603C6557445 @default.
- W4200126603 hasConceptScore W4200126603C71924100 @default.
- W4200126603 hasConceptScore W4200126603C86803240 @default.
- W4200126603 hasFunder F4320321001 @default.
- W4200126603 hasFunder F4320321133 @default.
- W4200126603 hasIssue "1" @default.
- W4200126603 hasLocation W42001266031 @default.
- W4200126603 hasOpenAccess W4200126603 @default.
- W4200126603 hasPrimaryLocation W42001266031 @default.
- W4200126603 hasRelatedWork W2961085424 @default.
- W4200126603 hasRelatedWork W3046775127 @default.
- W4200126603 hasRelatedWork W3047144510 @default.
- W4200126603 hasRelatedWork W3170094116 @default.
- W4200126603 hasRelatedWork W4205958290 @default.
- W4200126603 hasRelatedWork W4285260836 @default.
- W4200126603 hasRelatedWork W4286629047 @default.
- W4200126603 hasRelatedWork W4306321456 @default.