%0 Journal Article
%T 基于半监督模糊聚类的图像压缩研究<br>Research on Image Compression Based on Semi-Supervised Fuzzy Clustering
%A 王宇
%A 徐圣兵
%A 蔡炜
%A 夏泓禧
%A 黄永盛
%J Journal of Image and Signal Processing
%P 127-134
%@ 2325-6745
%D 2021
%I Hans Publishing
%R 10.12677/JISP.2021.103014
%X 针对主流图像压缩方法色彩深度冗余、无法针对特定场景加入标签以优化压缩质量等问题，本文提出了一种基于半监督模糊聚类(SFCM)的图像压缩方法，相比传统的图像压缩算法，该方法能通过引入模糊标签信息以提高特定应用场景下图像压缩的质量，即对特定区域进行标记以达到更好的压缩效果，从而使得图像压缩在更多不同应用场景保留更丰富的信息。本文实验选取传统的Lena图和COVID-19CT图像，实验结果显示此改进的图像压缩方法相比JPEG、K-Means等方法压缩得到的图像具有更好的信噪比。<br />
To solve the problems of image compression methods such as color redundancy and the inability to add label for specific scenes to gain compression quality, this paper proposes an image compression method based on semi-supervised fuzzy clustering (SFCM). Compared with traditional image compression algorithms, this method can improve the quality of image compression in specific scenarios by introducing fuzzy label, marking specific areas to achieve better compression effects, so that image compression retains more details in specific scenarios. The experiments using Lena images and COVID-19 CT images show that this improved image compression method has a better SNR than traditional methods such as JPEG and K-Means.
%K 半监督模糊聚类，图像压缩，标签信息<br>Semi-Supervised Fuzzy Clustering
%K Image Compression
%K Label Information
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=43860