FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2982166103> ?p ?o ?g. }

• W2982166103 abstract "Recently, deep learning-based saliency prediction methods have achieved significant accuracy improvements. However, they are hard to embed in practical multimedia applications due to large memory consumption and running time caused by complicated architectures. In addition, most methods are fine-tuned from pre-trained models for classification tasks, and networks cannot flexibly be transferred for a new task. In this paper, a condensed and randomly initialized student network is employed to achieve higher efficiency by transferring knowledge from complicated and well-trained teacher networks. This is the first use of knowledge distillation for efficient pixel-wise saliency prediction. Instead of directly minimizing Euclidean distance between feature maps, we propose two statistical representations of feature maps (i.e., first-order and second-order statistics) as knowledge. We conduct experiments on three kinds of teacher networks and four benchmark datasets to verify the effectiveness of the proposed method. Compared with the teacher networks, the student networks achieve an acceleration ratio of 4.56-4.73. Compared with state-of-the-art approaches, the proposed model achieves competitive accuracy with faster running speed (up to 4.38 times) and smaller model size (up to 93.27% reduction). We further embedded the proposed saliency prediction model into a video captioning application. The saliency-embedded approaches improve video captioning on all test metrics with a small complexity cost. The student-model embedded approach achieves 25% time saving with similar performance to the teacher embedded one." @default.
• W2982166103 created "2019-11-01" @default.
• W2982166103 creator A5007962086 @default.
• W2982166103 creator A5028597017 @default.
• W2982166103 creator A5029360035 @default.
• W2982166103 creator A5064564309 @default.
• W2982166103 creator A5068669736 @default.
• W2982166103 creator A5086252890 @default.
• W2982166103 creator A5088659027 @default.
• W2982166103 date "2019-10-15" @default.
• W2982166103 modified "2023-10-10" @default.
• W2982166103 title "Training Efficient Saliency Prediction Models with Knowledge Distillation" @default.
• W2982166103 cites W1942214758 @default.
• W2982166103 cites W1948751323 @default.
• W2982166103 cites W2005619762 @default.
• W2982166103 cites W2006396222 @default.
• W2982166103 cites W2035325201 @default.
• W2982166103 cites W2039313011 @default.
• W2982166103 cites W2043331342 @default.
• W2982166103 cites W2086791339 @default.
• W2982166103 cites W2133459682 @default.
• W2982166103 cites W2139501017 @default.
• W2982166103 cites W2194775991 @default.
• W2982166103 cites W2240167975 @default.
• W2982166103 cites W2294182682 @default.
• W2982166103 cites W2395611524 @default.
• W2982166103 cites W2507697802 @default.
• W2982166103 cites W2518599539 @default.
• W2982166103 cites W2527145521 @default.
• W2982166103 cites W2569272946 @default.
• W2982166103 cites W2592152820 @default.
• W2982166103 cites W2744613561 @default.
• W2982166103 cites W2750432752 @default.
• W2982166103 cites W2752191396 @default.
• W2982166103 cites W2765448945 @default.
• W2982166103 cites W2765881872 @default.
• W2982166103 cites W2794517705 @default.
• W2982166103 cites W2807746031 @default.
• W2982166103 cites W2884555738 @default.
• W2982166103 cites W2887585070 @default.
• W2982166103 cites W2895845501 @default.
• W2982166103 cites W2896034938 @default.
• W2982166103 cites W2896878184 @default.
• W2982166103 cites W2896897397 @default.
• W2982166103 cites W2963032190 @default.
• W2982166103 cites W2963112696 @default.
• W2982166103 cites W2963823251 @default.
• W2982166103 cites W2963906836 @default.
• W2982166103 cites W3103844505 @default.
• W2982166103 cites W3104979525 @default.
• W2982166103 cites W4239147634 @default.
• W2982166103 doi "https://doi.org/10.1145/3343031.3351089" @default.
• W2982166103 hasPublicationYear "2019" @default.
• W2982166103 type Work @default.
• W2982166103 sameAs 2982166103 @default.
• W2982166103 citedByCount "11" @default.
• W2982166103 countsByYear W29821661032020 @default.
• W2982166103 countsByYear W29821661032021 @default.
• W2982166103 countsByYear W29821661032022 @default.
• W2982166103 countsByYear W29821661032023 @default.
• W2982166103 crossrefType "proceedings-article" @default.
• W2982166103 hasAuthorship W2982166103A5007962086 @default.
• W2982166103 hasAuthorship W2982166103A5028597017 @default.
• W2982166103 hasAuthorship W2982166103A5029360035 @default.
• W2982166103 hasAuthorship W2982166103A5064564309 @default.
• W2982166103 hasAuthorship W2982166103A5068669736 @default.
• W2982166103 hasAuthorship W2982166103A5086252890 @default.
• W2982166103 hasAuthorship W2982166103A5088659027 @default.
• W2982166103 hasBestOaLocation W29821661031 @default.
• W2982166103 hasConcept C119857082 @default.
• W2982166103 hasConcept C121332964 @default.
• W2982166103 hasConcept C153294291 @default.
• W2982166103 hasConcept C154945302 @default.
• W2982166103 hasConcept C185592680 @default.
• W2982166103 hasConcept C204030448 @default.
• W2982166103 hasConcept C2777211547 @default.
• W2982166103 hasConcept C41008148 @default.
• W2982166103 hasConcept C43617362 @default.
• W2982166103 hasConceptScore W2982166103C119857082 @default.
• W2982166103 hasConceptScore W2982166103C121332964 @default.
• W2982166103 hasConceptScore W2982166103C153294291 @default.
• W2982166103 hasConceptScore W2982166103C154945302 @default.
• W2982166103 hasConceptScore W2982166103C185592680 @default.
• W2982166103 hasConceptScore W2982166103C204030448 @default.
• W2982166103 hasConceptScore W2982166103C2777211547 @default.
• W2982166103 hasConceptScore W2982166103C41008148 @default.
• W2982166103 hasConceptScore W2982166103C43617362 @default.
• W2982166103 hasFunder F4320306076 @default.
• W2982166103 hasFunder F4320322847 @default.
• W2982166103 hasLocation W29821661031 @default.
• W2982166103 hasOpenAccess W2982166103 @default.
• W2982166103 hasPrimaryLocation W29821661031 @default.
• W2982166103 hasRelatedWork W2961085424 @default.
• W2982166103 hasRelatedWork W3046775127 @default.
• W2982166103 hasRelatedWork W3107474891 @default.
• W2982166103 hasRelatedWork W3170094116 @default.
• W2982166103 hasRelatedWork W3209574120 @default.
• W2982166103 hasRelatedWork W4205958290 @default.
• W2982166103 hasRelatedWork W4286629047 @default.
• W2982166103 hasRelatedWork W4306321456 @default.

• next