کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
739092 | 1461624 | 2015 | 8 صفحه PDF | دانلود رایگان |
• A molecular hyperspectral imaging system is used to observe skin sections.
• Spectral angle and distance score are calculated for epithelial cells segmentation.
• The proposed method use spectral rather than spatial information for segmentation.
In the diagnosis of skin diseases by analyzing histopathological images of skin sections, the automated segmentation of cells in the epidermis area is an important step. Light microscopy based traditional methods usually cannot generate satisfying segmentation results due to complicated skin structures and limited information of this kind of image. In this study, we use a molecular hyperspectral imaging system to observe skin sections and propose a spectral based algorithm to segment epithelial cells. Unlike pixel-wise segmentation methods, the proposed algorithm considers both the spectral angle and the distance score between the test and the reference spectrum for segmentation. The experimental results indicate that the proposed algorithm performs better than the K-Means, fuzzy C-means, and spectral angle mapper algorithms because it can identify pixels with similar spectral angle but a different spectral distance.
Journal: Optics & Laser Technology - Volume 74, November 2015, Pages 79–86