English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The use of multiple and semantically correlated sources can providecomplementary information to each other that may not be evident when workingwith individual modalities on their own. In this context, multi-modal modelscan help producing more accurate and robust predictions in machine learningtasks where audio-visual data is available. This paper presents a multi-modalmodel for automatic scene classification that exploits simultaneously auditoryand visual information. The proposed approach makes use of two separatenetworks which are respectively trained in isolation on audio and visual data,so that each network specializes in a given modality. The visual subnetwork isa pre-trained VGG16 model followed by a bidiretional recurrent layer, while theresidual audio subnetwork is based on stacked squeeze-excitation convolutionalblocks trained from scratch. After training each subnetwork, the fusion ofinformation from the audio and visual streams is performed at two differentstages. The early fusion stage combines features resulting from the lastconvolutional block of the respective subnetworks at different time steps tofeed a bidirectional recurrent structure. The late fusion stage combines theoutput of the early fusion stage with the independent predictions provided bythe two subnetworks, resulting in the final prediction. We evaluate the methodusing the recently published TAU Audio-Visual Urban Scenes 2021, which containssynchronized audio and video recordings from 12 European cities in 10 differentscene classes. The proposed model has been shown to provide an excellenttrade-off between prediction performance (86.5%) and system complexity (15Mparameters) in the evaluation results of the DCASE 2021 Challenge.

Squeeze-Excitation Convolutional Recurrent Neural Networks for Audio-Visual Scene Classification