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...

Popoln opis

Bibliografske podrobnosti

Principais autores:	Peng Yang, Chunmei Li, Chengwu Fang, Shasha Kong, Yunpeng Jin, Kai Li, Haiyang Li, Xiangjie Huang, Yaosheng Han
Format:	Artigo
Jezik:	Inglês
Izdano:	IEEE 2024-01-01
Serija:	IEEE Access
Teme:	Deep learning semantic segmentation multi-scale context semantic context pixel contrast
Online dostop:	https://ieeexplore.ieee.org/document/10456764/
Oznake:	Označite Brez oznak, prvi označite!