FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2989763126> ?p ?o ?g. }

• W2989763126 abstract "In late fusion, each modality is processed in a separate unimodal Convolutional Neural Network (CNN) stream and the scores of each modality are fused at the end. Due to its simplicity late fusion is still the predominant approach in many state-of-the-art multimodal applications. In this paper, we present a simple neural network module for leveraging the knowledge from multiple modalities in convolutional neural networks. The propose unit, named Multimodal Transfer Module (MMTM), can be added at different levels of the feature hierarchy, enabling slow modality fusion. Using squeeze and excitation operations, MMTM utilizes the knowledge of multiple modalities to recalibrate the channel-wise features in each CNN stream. Despite other intermediate fusion methods, the proposed module could be used for feature modality fusion in convolution layers with different spatial dimensions. Another advantage of the proposed method is that it could be added among unimodal branches with minimum changes in the their network architectures, allowing each branch to be initialized with existing pretrained weights. Experimental results show that our framework improves the recognition accuracy of well-known multimodal networks. We demonstrate state-of-the-art or competitive performance on four datasets that span the task domains of dynamic hand gesture recognition, speech enhancement, and action recognition with RGB and body joints." @default.
• W2989763126 created "2019-12-05" @default.
• W2989763126 creator A5022551740 @default.
• W2989763126 creator A5030257576 @default.
• W2989763126 creator A5032931538 @default.
• W2989763126 creator A5084879161 @default.
• W2989763126 date "2019-11-20" @default.
• W2989763126 modified "2023-09-27" @default.
• W2989763126 title "MMTM: Multimodal Transfer Module for CNN Fusion" @default.
• W2989763126 cites W1518556865 @default.
• W2989763126 cites W1522734439 @default.
• W2989763126 cites W1599510591 @default.
• W2989763126 cites W1755205674 @default.
• W2989763126 cites W1836465849 @default.
• W2989763126 cites W1882423120 @default.
• W2989763126 cites W1943054406 @default.
• W2989763126 cites W1947481528 @default.
• W2989763126 cites W1993666393 @default.
• W2989763126 cites W2009059481 @default.
• W2989763126 cites W2016053056 @default.
• W2989763126 cites W2024051019 @default.
• W2989763126 cites W2031767521 @default.
• W2989763126 cites W2052666245 @default.
• W2989763126 cites W2070707809 @default.
• W2989763126 cites W2108598243 @default.
• W2989763126 cites W2141939040 @default.
• W2989763126 cites W2141998673 @default.
• W2989763126 cites W2150884987 @default.
• W2989763126 cites W2154754161 @default.
• W2989763126 cites W2156303437 @default.
• W2989763126 cites W2159202424 @default.
• W2989763126 cites W2184188583 @default.
• W2989763126 cites W2194775991 @default.
• W2989763126 cites W2342311830 @default.
• W2989763126 cites W2471695703 @default.
• W2989763126 cites W2525648609 @default.
• W2989763126 cites W2526449353 @default.
• W2989763126 cites W2559809918 @default.
• W2989763126 cites W2587372848 @default.
• W2989763126 cites W2587989515 @default.
• W2989763126 cites W2595328592 @default.
• W2989763126 cites W2596627958 @default.
• W2989763126 cites W2607014458 @default.
• W2989763126 cites W2619947201 @default.
• W2989763126 cites W2728829052 @default.
• W2989763126 cites W2746301562 @default.
• W2989763126 cites W2767290858 @default.
• W2989763126 cites W2769039400 @default.
• W2989763126 cites W2770472008 @default.
• W2989763126 cites W2781400102 @default.
• W2989763126 cites W2788241093 @default.
• W2989763126 cites W2788865504 @default.
• W2989763126 cites W2789011059 @default.
• W2989763126 cites W2791526950 @default.
• W2989763126 cites W2797090057 @default.
• W2989763126 cites W2798498022 @default.
• W2989763126 cites W2799211965 @default.
• W2989763126 cites W2808631503 @default.
• W2989763126 cites W2888471892 @default.
• W2989763126 cites W2894879265 @default.
• W2989763126 cites W2895368507 @default.
• W2989763126 cites W2901702433 @default.
• W2989763126 cites W2903314716 @default.
• W2989763126 cites W2904106524 @default.
• W2989763126 cites W2951183276 @default.
• W2989763126 cites W2951805656 @default.
• W2989763126 cites W2956679704 @default.
• W2989763126 cites W2962835968 @default.
• W2989763126 cites W2962960500 @default.
• W2989763126 cites W2963082324 @default.
• W2989763126 cites W2963155035 @default.
• W2989763126 cites W2963180222 @default.
• W2989763126 cites W2963282966 @default.
• W2989763126 cites W2963307811 @default.
• W2989763126 cites W2963369114 @default.
• W2989763126 cites W2963420686 @default.
• W2989763126 cites W2963524571 @default.
• W2989763126 cites W2963681914 @default.
• W2989763126 cites W2963717374 @default.
• W2989763126 cites W304179306 @default.
• W2989763126 cites W3123318516 @default.
• W2989763126 cites W914561379 @default.
• W2989763126 hasPublicationYear "2019" @default.
• W2989763126 type Work @default.
• W2989763126 sameAs 2989763126 @default.
• W2989763126 citedByCount "0" @default.
• W2989763126 crossrefType "posted-content" @default.
• W2989763126 hasAuthorship W2989763126A5022551740 @default.
• W2989763126 hasAuthorship W2989763126A5030257576 @default.
• W2989763126 hasAuthorship W2989763126A5032931538 @default.
• W2989763126 hasAuthorship W2989763126A5084879161 @default.
• W2989763126 hasConcept C138885662 @default.
• W2989763126 hasConcept C144024400 @default.
• W2989763126 hasConcept C153180895 @default.
• W2989763126 hasConcept C154945302 @default.
• W2989763126 hasConcept C158525013 @default.
• W2989763126 hasConcept C2776401178 @default.
• W2989763126 hasConcept C2779903281 @default.
• W2989763126 hasConcept C2780226545 @default.
• W2989763126 hasConcept C36289849 @default.

• next