WMSA–WBS: Efficient Wave Multi-Head Self-Attention with Wavelet Bottleneck

The critical component of the vision transformer (ViT) architecture is multi-head self-attention (MSA), which enables the encoding of long-range dependencies and heterogeneous interactions. However, MSA has two significant limitations: its limited ability to capture local features and its high compu...

Descripción completa

Detalles Bibliográficos

Autores principales:	Xiangyang Li, Yafeng Li, Pan Fan, Xueya Zhang
Formato:	Artigo
Lenguaje:	Inglês
Publicado:	MDPI AG 2025-08-01
Colección:	Sensors
Materias:	vision transformer self-attention waveflow wavelet transform convolution neural network
Acceso en línea:	https://www.mdpi.com/1424-8220/25/16/5046
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!