MSGFormer: A DeepLabv3+ Like Semantically Masked and Pixel Contrast Transformer for MouseHole Segmentation

In semantic segmentation, the efficient representation of multi-scale context is of paramount importance. Inspired by the remarkable performance of Vision Transformers (ViT) in image classification, subsequent researchers have proposed some Semantic Segmentation ViTs, most of which have achieved imp...

ver descrição completa

Detalhes bibliográficos

Principais autores:	Peng Yang, Chunmei Li, Chengwu Fang, Shasha Kong, Yunpeng Jin, Kai Li, Haiyang Li, Xiangjie Huang, Yaosheng Han
Formato:	Artigo
Idioma:	Inglês
Publicado em:	IEEE 2024-01-01
coleção:	IEEE Access
Assuntos:	Deep learning semantic segmentation multi-scale context semantic context pixel contrast
Acesso em linha:	https://ieeexplore.ieee.org/document/10456764/
Tags:	Adicionar Tag Sem tags, seja o primeiro a adicionar uma tag!