基于CiteSpace的抠图研究热点与前沿分析

Hotspots and Cutting-edge Analysis of Matting Based on CiteSpace

为了对抠图发展领域进行深入的研究,全面检索中国知网从2000年以来关于抠图的文献,利用CiteSpace文献计量学工具绘制知识图谱并分析。结果显示,抠图研究的文献发文量总体趋势相对不稳定,呈略微下降趋势;抠图领域作者的发文量普遍较少,各发文机构之间并没有形成密切的合作关系;关键词分析显示抠图、通道、图像处理和图像抠图为该领域的研究热点,到2020年之后,现有的抠图领域研究成果非常稀疏;此外,深度学习、深度神经网络、注意力机制和人像抠图是抠图领域未来的研究趋势。

This paper comprehensively searches the literature on cutouts from CNKI since 2000,and uses CiteSpace bibliometrics tools to draw and analyze knowledge maps.The results showed that the overall trend of the number of papers published in the matting study was relatively unstable,showing a slight downward trend.The number of authors in the field of matting is generally small,and two close-knit teams have been formed,and there is no close cooperative relationship between the publishing institutions.Keyword analysis shows that matting,channel,image processing and image matting are the research hotspots in this field,and after 2020,the existing research results in the field of matting are very sparse.In addition,deep learning,deep neural networks,attention mechanism and portrait matting are the future research trends in the field of matting.