KEYWORDS: Skin, Skin cancer, Feature extraction, Data processing, Principal component analysis, In vivo imaging, Diffuse reflectance spectroscopy, Autofluorescence
Non-invasive diagnosis of skin pathologies as skin cancer using optical methods has become increasingly common in recent years. However, the related skin data processing is often quite complex, and the way in which this step is carried out can significantly affect the final results. This study presents the results of diffuse reflectance spectra (with spectral range of the emission source is 300-800 nm) and autofluorescence spectra (with 7 autofluorescence excitation wavelengths in the 360-430 nm range) obtained in vivo from precancerous and benign skin lesions of various types (compensatory hyperplasia, atypical hyperplasia and dysplasia). The skin lesions were modelled using a preclinical model in mice. Spectra were taken in the range of 317 - 789 nm at three different source-detector separations: 271, 536 and 834 μm. The spectra obtained were processed using statistical feature extraction techniques, traditional machine learning (support vector machine, linear discriminant analysis, k-nearest neighbors) and deep learning methods (artificial neural network, convolutional neural network, autoencoder). This study presents a comparison of the performance of these methods and their combinations for multiclass classification of skin lesions.
Optical biopsy methods, which consists of analysing the response of tissue to light excitation, are being increasingly used in recent years for the diagnosis of skin pathologies. At the same time, the use of multimodal methods often significantly increases diagnostic efficiency as well as extending the limits of applicability of the methods. This contribution presents the results of in vivo analysis of precancerous and benign skin conditions (compensatory hyperplasia, atypical hyperplasia and dysplasia) in mice preclinical model, based on bimodal spectroscopic data, including multiply excited autofluorescence with 7 autofluorescence excitation wavelengths in the 360-430 nm range and diffuse reflectance spectroscopy with xenon lamp, that emits mainly in the 300-800 nm spectral range, as a source. The instrument used in this study provided the ability to collect spectra in the spectral range 317 - 789 nm at three different source-detector separations: 271, 536 and 834 μm. The results were processed using machine learning methods (principal component analysis, support vector machine, linear discriminant analysis, artificial neural network) and then various data fusion methods (Stacking, Begging, Boosting, Voting) were implemented to combine the results of analysis of all the modalities. This study presents a comparison of the performance of these data fusion methods. The results obtained in this work can be further applied to the diagnosis of carcinoma using optical biopsy methods.
The aim of the current study is to evaluate the classification accuracy and provide corresponding biological interpretation of four classification methods used on autofluorescence (AF) and diffuse reflectance (DR) spectra acquired in vivo on healthy human skin of different phototypes, civil and apparent age groups. Spectroscopic data were acquired on 91 patients using the SpectroLive device. The latter spatially and spectrally-resolved device features four source-to-detector distances (D1-D4) and six excitation light sources: 5 peaks for AF and one broadband white light for DR. For all patients, spectra were acquired on two healthy skin sites i.e. hand palm and inner wrist chosen for their low sun exposure. Four classification methods were tested: Support Vector Machine, K-Nearest Neighbors, Linear Discriminant Analysis and Artificial Neural Network. All combinations of excitation wavelengths, distances and skin sites acquisition were tested to find out the best classification results following a training step on 67 % of the dataset and a validation step on 33 % of the dataset. Classification accuracies were compared using Principal Components Analysis and statistical features. For civil and biological skin age groups discrimination, best classification results (70 % and 76 % respectively) were obtained when combining autofluorescence spectral features from three excitation wavelengths (385, 395 and 405 nm) all acquired at the shortest distance (400 µm) on hand palm. The combination of AF, inner wrist and the longest distance (1 mm) gave the best classification results (76 %) for phototype groups discrimination.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.