受损皮肤屏障的皮肤镜图像相关特征  

Dermoscopic Image Characteristics Related to Damaged Skin Barriers

作　　者：陈逸轩 何冰冰 郭振宇 张梅 张宁涛 张榆锋[1] CHEN Yi-xuan;HE Bing-bing;GUO Zhen-yu;ZHANG Mei;ZHANG Ning-tao;ZHANG Yu-feng(School of Information,Yunnan University,Kunming 650091,China;Yunnan Betani Biotechnology Group Co.,Ltd.,Kunming 650091,China)

机构地区：[1]云南大学信息学院,昆明650091 [2]云南贝泰妮生物科技集团股份有限公司,昆明650091

出　　处：《科学技术与工程》2024年第34期14748-14753,共6页Science Technology and Engineering

基　　金：国家自然科学基金(62261057);云南省2022年博士后科研基金研究项目(202210563)。

摘　　要：准确评估皮肤屏障受损程度对指导皮肤相关的护理和治疗具有重要意义。经皮水分散失(trans epidermal water loss, TEWL)是评估皮肤屏障受损的金标准,但其测量复杂且成本高。皮肤镜图像具有简便、低价、无创、实时等优点。为了研究受损皮肤屏障的皮肤镜图像相关特征,首先对皮肤镜图像进行白平衡、灰度化、直方图均衡化及维纳滤波预处理,之后提取图像的统计特征、纹理特征和边缘特征,并与皮肤受损造模小鼠的TEWL检测结果进行相关性分析。结果表明,纹理特征与TEWL值之间相关性较高,特别半径为1的局部二值特征的熵相关性最高,相关系数为0.752。皮肤镜的纹理特征能较好刻画皮肤屏障受损程度,有助于替代TEWL开发皮肤屏障功能评价的新方法。The accurate assessment of skin barrier impairment is of significant importance in guiding skin-related care and treatment.TEWL(trans epidermal water loss)is considered the gold standard for evaluating skin barrier damage,but its measurement is complex and costly.Dermoscopy images offer advantages such as simplicity,low cost,non-invasiveness,and real-time capabilities.The features of dermoscopy images related to impaired skin barrier were investigated.Dermoscopy images were initially preprocessed using white balance,grayscaling,histogram equalization,and Wiener filtering.Subsequently,statistical features,texture features,and edge features of the images were extracted and analyzed for correlation with TEWL detection results in a skin-damaged mouse model.The results indicate a high correlation between texture features and TEWL values,especially the entropy of local binary patterns with a radius of 1,which has the highest correlation coefficient of 0.752.The texture features of dermoscopy images are found to effectively depict the degree of skin barrier damage,suggesting their potential in developing new methods for skin barrier function assessment as an alternative to TEWL.

关 键 词：皮肤屏障 经皮水分散失 皮肤镜图像 图像纹理特征 相关性分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]