基于Transformer的多维度图像应用  

Multi-dimensional Image Application Based on Transformer

作　　者：高政霞 GAO Zhengxia

机构地区：[1]兰州资源环境职业技术大学信息工程学院,兰州730021

出　　处：《科技创新与应用》2023年第10期5-8,共4页Technology Innovation and Application

基　　金：工信行指委2022—2023年度科研课题项目(工信行指委[2022]4号,GXHZWC82141);中国高校产学研创新基金-北创助教项目(二期)(2021BCE02006);兰州资源环境职业技术大学教学改革研究课题项目(26)。

摘　　要：Transformer自提出以来便热度不降,特别是在自然语言处理领域热度不减。该文首先回顾Transformer的发展历程,分析Transformer在自然语言处理领域的特点和优缺点。总结基于计算机视觉领域应用Transformer模型的技术瓶颈,特别是基于Transformer的多维度图像分类的模型应用C-Tran模型,该文着重于Transformer对计算机视觉领域的革新和模型应用,在现有研究基础上分析多标签多维度图像应用的模型及模型效果,最后提出多维度视觉图像的具体应用情况。Transformer has been popular since it was proposed,especially in the field of natural language processing.Firstly,this paper reviews the development of Transformer,analyzes the characteristics,advantages and disadvantages of Transformer in the field of natural language processing,and summarizes the technical bottleneck of the application of Transformer model in the field of computer vision,especially the application of C-Tran model in the model of multi-dimensional image classification based on Transformer.Then,this paper focuses on the innovation and model application of Transformer in the field of computer vision,based on the existing research,the application model and model effect of multi-label multidimensional image are analyzed.Finally,the specific application of multidimensional visual image is proposed.

关 键 词：TRANSFORMER 多标签 多维度图像分类 C-Tran 应用 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]