当前位置: X-MOL 学术IEEE Sens. J. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Multifractal Texture Analysis of Salivary Fern Pattern for Oral Pre-Cancers and Cancer Assessment
IEEE Sensors Journal ( IF 4.3 ) Pub Date : 2021-01-21 , DOI: 10.1109/jsen.2021.3053262
Neha Sharma , Debaleena Nawn , Sawon Pratiher , Sayani Shome , Ritam Chatterjee , Karabi Biswas , Mousumi Pal , Ranjan Rashmi Paul , Srimonti Dutta , Jyotirmoy Chatterjee

Saliva has emerged as an efficient screening sample for early stage detection of oral cancer (OC) owing to non-invasiveness coupled with high sensitivity and specificity. Although spectroscopic characterization of saliva in oral potentially malignant disorders OPMDs) and OC is extensively studied, its potential as imaging biomarker is sparsely explored. Further, the literature on crystalline pattern of saliva for other diseases or different physiological conditions is mostly qualitative. This paper proposed multifractal based methodology to quantitatively study alteration of the salivary fern pattern in different OPMDs and OC in relation to normal counterpart. The fern pattern of dried saliva is captured by stereo-zoom microscope in reflective mode and an image dataset is developed. We resort to two dimensional multi-fractal detrended fluctuation analysis (2d MFDFA) to elucidate the complexity and heterogeneity of these micro-structured patterns. Existence of multifractal nature embedded in salivary fern has been validated for the first time. Long range spatial correlation is found to be the origin of multifractality. Variation in multi-scale self-similarity of irregular pattern in different study groups is demonstrated by four features extracted from MFDFA. Statistical analysis shows discriminating nature of these features for combinations of pairwise interclass classification. This study sheds light on acceptability of microscopic images of arborized saliva in fast and cost effective screening of different oral lesions.

中文翻译:

口腔癌前唾液蕨模式的多重分形纹理分析和癌症评估

由于无创性以及高灵敏度和特异性,唾液已成为早期检测口腔癌(OC)的有效筛选样品。尽管对口腔潜在恶性疾病(OPMDs)和OC中唾液的光谱特性进行了广泛研究,但仍在稀疏地探索其作为成像生物标志物的潜力。此外,有关其他疾病或不同生理状况的唾液结晶型的文献大多是定性的。本文提出了基于多重分形的方法,定量研究了不同OPMDs和OC中唾液蕨模式相对于正常唾液的变化。通过立体变焦显微镜以反射模式捕获干燥唾液的蕨模式,并开发图像数据集。我们诉诸二维多维分形趋势波动分析(二维MFDFA),以阐明这些微观结构模式的复杂性和异质性。唾液蕨中嵌入的多重分形性质的存在已得到首次验证。发现远距离空间相关性是多重分形的起源。从MFDFA中提取的四个特征证明了不同研究组中不规则模式的多尺度自相似性的变化。统计分析表明,这些特征对于成对类间分类的组合具有区别性。这项研究揭示了在快速且经济有效地筛查不同口腔病变的情况下,人工唾液显微图像的可接受性。唾液蕨中嵌入的多重分形性质的存在已得到首次验证。发现远距离空间相关性是多重分形的起源。从MFDFA中提取的四个特征证明了不同研究组中不规则模式的多尺度自相似性的变化。统计分析表明,这些特征对于成对类间分类的组合具有区别性。这项研究揭示了在快速且经济有效地筛查不同口腔病变的情况下,人工唾液显微图像的可接受性。唾液蕨中嵌入的多重分形性质的存在已得到首次验证。发现远距离空间相关性是多重分形的起源。从MFDFA中提取的四个特征证明了不同研究组中不规则模式的多尺度自相似性的变化。统计分析表明,这些特征对于成对类间分类的组合具有区别性。这项研究揭示了在快速且经济有效地筛查不同口腔病变的情况下,人工唾液显微图像的可接受性。从MFDFA中提取的四个特征证明了不同研究组中不规则模式的多尺度自相似性的变化。统计分析表明,这些特征对于成对类间分类的组合具有区别性。这项研究揭示了在快速且经济有效地筛查不同口腔病变的情况下,人工唾液显微图像的可接受性。从MFDFA中提取的四个特征证明了不同研究组中不规则模式的多尺度自相似性的变化。统计分析表明,这些特征对于成对类间分类的组合具有区别性。这项研究揭示了在快速且经济有效地筛查不同口腔病变的情况下,人工唾液显微图像的可接受性。
更新日期:2021-03-05
down
wechat
bug