Optical Coherence Tomography (OCT) offers real-time high-resolution three-dimensional images of tissue microstructures. In this study, we used OCT skin images acquired from ten volunteers, neither of whom had any skin conditions addressing the features of their anatomic location. OCT segmented images are analyzed based on their optical properties (attenuation coefficient) and textural image features e.g., contrast, correlation, homogeneity, energy, entropy, etc. Utilizing the information and referring to their clinical insight, we aim to make a comprehensive computational model for the healthy skin. The derived parameters represent the OCT microstructural morphology and might provide biological information for generating an atlas of normal skin from different anatomic sites of human skin and may allow for identification of cell microstructural changes in cancer patients. We then compared the parameters of healthy samples with those of abnormal skin and classified them using a linear Support Vector Machines (SVM) with 82% accuracy.
Adabi, S., Conforto, S., Hosseinzadeh, M., Noe, S., Daveluy, S., Mehregan, D., et al. (2017). Textural analysis of optical coherence tomography skin images: Quantitative differentiation between healthy and cancerous tissues. In Progress in Biomedical Optics and Imaging - Proceedings of SPIE (pp.100533F). SPIE [10.1117/12.2254869].
Textural analysis of optical coherence tomography skin images: Quantitative differentiation between healthy and cancerous tissues
ADABI, SABA;CONFORTO, SILVIA;
2017-01-01
Abstract
Optical Coherence Tomography (OCT) offers real-time high-resolution three-dimensional images of tissue microstructures. In this study, we used OCT skin images acquired from ten volunteers, neither of whom had any skin conditions addressing the features of their anatomic location. OCT segmented images are analyzed based on their optical properties (attenuation coefficient) and textural image features e.g., contrast, correlation, homogeneity, energy, entropy, etc. Utilizing the information and referring to their clinical insight, we aim to make a comprehensive computational model for the healthy skin. The derived parameters represent the OCT microstructural morphology and might provide biological information for generating an atlas of normal skin from different anatomic sites of human skin and may allow for identification of cell microstructural changes in cancer patients. We then compared the parameters of healthy samples with those of abnormal skin and classified them using a linear Support Vector Machines (SVM) with 82% accuracy.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
