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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In this paper, we show that ImageNet-Pretrained standard deep CNN models canbe used as strong baseline networks for audio classification. Even though thereis a significant difference between audio Spectrogram and standard ImageNetimage samples, transfer learning assumptions still hold firmly. To understandwhat enables the ImageNet pretrained models to learn useful audiorepresentations, we systematically study how much of pretrained weights isuseful for learning spectrograms. We show (1) that for a given standard modelusing pretrained weights is better than using randomly initialized weights (2)qualitative results of what the CNNs learn from the spectrograms by visualizingthe gradients. Besides, we show that even though we use the pretrained modelweights for initialization, there is variance in performance in various outputruns of the same model. This variance in performance is due to the randominitialization of linear classification layer and random mini-batch orderingsin multiple runs. This brings significant diversity to build stronger ensemblemodels with an overall improvement in accuracy. An ensemble of ImageNetpretrained DenseNet achieves 92.89% validation accuracy on the ESC-50 datasetand 87.42% validation accuracy on the UrbanSound8K dataset which is the currentstate-of-the-art on both of these datasets.

Rethinking CNN Models for Audio Classification