FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4319003893> ?p ?o ?g. }

• W4319003893 endingPage "732" @default.
• W4319003893 startingPage "732" @default.
• W4319003893 abstract "Many applications, such as autonomous driving, robotics, etc., require accurately estimating depth in real time. Currently, deep learning is the most popular approach to stereo depth estimation. Some of these models have to operate in highly energy-constrained environments, while they are usually computationally intensive, containing massive parameter sets ranging from thousands to millions. This makes them hard to perform on low-power devices with limited storage in practice. To overcome this shortcoming, we model the training process of a deep neural network (DNN) for depth estimation under a given energy constraint as a constrained optimization problem and solve it through a proposed projected adaptive cubic quasi-Newton method (termed ProjACQN). Moreover, the trained model is also deployed on a GPU and an embedded device to evaluate its performance. Experiments show that the stage four results of ProjACQN on the KITTI-2012 and KITTI-2015 datasets under a 70% energy budget achieve (1) 0.13% and 0.61%, respectively, lower three-pixel error than the state-of-the-art ProjAdam when put on a single RTX 3090Ti; (2) 4.82% and 7.58%, respectively, lower three-pixel error than the pruning method Lottery-Ticket; (3) 5.80% and 0.12%, respectively, lower three-pixel error than ProjAdam on the embedded device Nvidia Jetson AGX Xavier. These results show that our method can reduce the energy consumption of depth estimation DNNs while maintaining their accuracy." @default.
• W4319003893 created "2023-02-03" @default.
• W4319003893 creator A5021293751 @default.
• W4319003893 creator A5025189927 @default.
• W4319003893 creator A5078001039 @default.
• W4319003893 creator A5084070365 @default.
• W4319003893 date "2023-02-01" @default.
• W4319003893 modified "2023-09-26" @default.
• W4319003893 title "Energy-Constrained Deep Neural Network Compression for Depth Estimation" @default.
• W4319003893 cites W1921093919 @default.
• W4319003893 cites W2017369466 @default.
• W4319003893 cites W2150066425 @default.
• W4319003893 cites W2795023265 @default.
• W4319003893 cites W2954174912 @default.
• W4319003893 cites W2962851801 @default.
• W4319003893 cites W2967115342 @default.
• W4319003893 cites W2968164956 @default.
• W4319003893 cites W2968826970 @default.
• W4319003893 cites W3011733903 @default.
• W4319003893 cites W3033569682 @default.
• W4319003893 cites W3034976873 @default.
• W4319003893 cites W3135576287 @default.
• W4319003893 cites W3159126906 @default.
• W4319003893 cites W3159270659 @default.
• W4319003893 cites W4210790836 @default.
• W4319003893 cites W4221099740 @default.
• W4319003893 cites W4283584216 @default.
• W4319003893 doi "https://doi.org/10.3390/electronics12030732" @default.
• W4319003893 hasPublicationYear "2023" @default.
• W4319003893 type Work @default.
• W4319003893 citedByCount "0" @default.
• W4319003893 crossrefType "journal-article" @default.
• W4319003893 hasAuthorship W4319003893A5021293751 @default.
• W4319003893 hasAuthorship W4319003893A5025189927 @default.
• W4319003893 hasAuthorship W4319003893A5078001039 @default.
• W4319003893 hasAuthorship W4319003893A5084070365 @default.
• W4319003893 hasBestOaLocation W43190038931 @default.
• W4319003893 hasConcept C105795698 @default.
• W4319003893 hasConcept C108010975 @default.
• W4319003893 hasConcept C108583219 @default.
• W4319003893 hasConcept C111919701 @default.
• W4319003893 hasConcept C113775141 @default.
• W4319003893 hasConcept C11413529 @default.
• W4319003893 hasConcept C119599485 @default.
• W4319003893 hasConcept C121332964 @default.
• W4319003893 hasConcept C127413603 @default.
• W4319003893 hasConcept C154945302 @default.
• W4319003893 hasConcept C160633673 @default.
• W4319003893 hasConcept C163258240 @default.
• W4319003893 hasConcept C186370098 @default.
• W4319003893 hasConcept C2776036281 @default.
• W4319003893 hasConcept C2780165032 @default.
• W4319003893 hasConcept C33923547 @default.
• W4319003893 hasConcept C41008148 @default.
• W4319003893 hasConcept C50644808 @default.
• W4319003893 hasConcept C62520636 @default.
• W4319003893 hasConcept C6557445 @default.
• W4319003893 hasConcept C78519656 @default.
• W4319003893 hasConcept C79403827 @default.
• W4319003893 hasConcept C86803240 @default.
• W4319003893 hasConcept C98045186 @default.
• W4319003893 hasConceptScore W4319003893C105795698 @default.
• W4319003893 hasConceptScore W4319003893C108010975 @default.
• W4319003893 hasConceptScore W4319003893C108583219 @default.
• W4319003893 hasConceptScore W4319003893C111919701 @default.
• W4319003893 hasConceptScore W4319003893C113775141 @default.
• W4319003893 hasConceptScore W4319003893C11413529 @default.
• W4319003893 hasConceptScore W4319003893C119599485 @default.
• W4319003893 hasConceptScore W4319003893C121332964 @default.
• W4319003893 hasConceptScore W4319003893C127413603 @default.
• W4319003893 hasConceptScore W4319003893C154945302 @default.
• W4319003893 hasConceptScore W4319003893C160633673 @default.
• W4319003893 hasConceptScore W4319003893C163258240 @default.
• W4319003893 hasConceptScore W4319003893C186370098 @default.
• W4319003893 hasConceptScore W4319003893C2776036281 @default.
• W4319003893 hasConceptScore W4319003893C2780165032 @default.
• W4319003893 hasConceptScore W4319003893C33923547 @default.
• W4319003893 hasConceptScore W4319003893C41008148 @default.
• W4319003893 hasConceptScore W4319003893C50644808 @default.
• W4319003893 hasConceptScore W4319003893C62520636 @default.
• W4319003893 hasConceptScore W4319003893C6557445 @default.
• W4319003893 hasConceptScore W4319003893C78519656 @default.
• W4319003893 hasConceptScore W4319003893C79403827 @default.
• W4319003893 hasConceptScore W4319003893C86803240 @default.
• W4319003893 hasConceptScore W4319003893C98045186 @default.
• W4319003893 hasFunder F4320321001 @default.
• W4319003893 hasFunder F4320322843 @default.
• W4319003893 hasIssue "3" @default.
• W4319003893 hasLocation W43190038931 @default.
• W4319003893 hasOpenAccess W4319003893 @default.
• W4319003893 hasPrimaryLocation W43190038931 @default.
• W4319003893 hasRelatedWork W2006342414 @default.
• W4319003893 hasRelatedWork W2072862869 @default.
• W4319003893 hasRelatedWork W2159159552 @default.
• W4319003893 hasRelatedWork W2168310718 @default.
• W4319003893 hasRelatedWork W2405058345 @default.
• W4319003893 hasRelatedWork W2599766002 @default.
• W4319003893 hasRelatedWork W2899726612 @default.

• next