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 work, we present DeFTAN-II, an efficient multichannel speechenhancement model based on transformer architecture and subgroup processing.Despite the success of transformers in speech enhancement, they face challengesin capturing local relations, reducing the high computational complexity, andlowering memory usage. To address these limitations, we introduce subgroupprocessing in our model, combining subgroups of locally emphasized featureswith other subgroups containing original features. The subgroup processing isimplemented in several blocks of the proposed network. In the proposed splitdense blocks extracting spatial features, a pair of subgroups is sequentiallyconcatenated and processed by convolution layers to effectively reduce thecomputational complexity and memory usage. For the F- and T-transformersextracting temporal and spectral relations, we introduce cross-attentionbetween subgroups to identify relationships between locally emphasized andnon-emphasized features. The dual-path feedforward network then aggregatesattended features in terms of the gating of local features processed by dilatedconvolutions. Through extensive comparisons with state-of-the-art multichannelspeech enhancement models, we demonstrate that DeFTAN-II with subgroupprocessing outperforms existing methods at significantly lower computationalcomplexity. Moreover, we evaluate the model's generalization capability onreal-world data without fine-tuning, which further demonstrates itseffectiveness in practical scenarios.

DeFTAN-II: Efficient Multichannel Speech Enhancement with Subgroup Processing