Fully Automated Identification of Skin Morphology in Raster-Scan Optoacoustic Mesoscopy using Artificial Intelligence.

PURPOSE Identification of morphological characteristics of skin lesions is of vital importance in diagnosing diseases with dermatological manifestations. This task is often performed manually or in an automated way based on intensity level. Recently, ultra-broadband raster-scan optoacoustic mesoscopy (UWB-RSOM) was developed to offer unique cross-sectional optical imaging of the skin. A machine learning (ML) approach is proposed here to enable, for the first time, automated identification of skin layers in UWB-RSOM data. METHODS AND MATERIALS The proposed method, termed SkinSeg, was applied to coronal UWB-RSOM images obtained from 12 human participants. SkinSeg is a multi-step methodology that integrates data processing and transformation, feature extraction, feature selection and classification. Various image features and learning models were tested for their suitability at discriminating skin layers including traditional machine learning along with more advanced deep learning algorithms. An SVM-based post-processing approach was finally applied to further improve the classification outputs. RESULTS Random forest proved to be the most effective technique, achieving mean classification accuracy of 86.89% evaluated based on a repeated leave-one-out strategy. Insights about the features extracted and their effect on classification accuracy are provided. The highest accuracy was achieved using a small group of 4 features and remained at the same level or was even slightly decreased when more features were included. Convolutional neural networks provided also promising results at a level of approximately 85%. The application of the proposed post-processing technique was proved to be effective in terms of both testing accuracy and 3D visualization of classification maps. CONCLUSIONS SkinSeg demonstrated unique potential in identifying skin layers. The proposed method may facilitate clinical evaluation, monitoring and diagnosis of diseases linked to skin inflammation, diabetes and skin cancer. This article is protected by copyright. All rights reserved.

[1]  Yoav Freund,et al.  A decision-theoretic generalization of on-line learning and an application to boosting , 1997, EuroCOLT.

[2]  Yiannis Aloimonos,et al.  LightNet: A Versatile, Standalone Matlab-based Environment for Deep Learning , 2016, ACM Multimedia.

[3]  V. Ntziachristos,et al.  Precision assessment of label-free psoriasis biomarkers with ultra-broadband optoacoustic mesoscopy , 2017, Nature Biomedical Engineering.

[4]  A Nisbet,et al.  The role of texture analysis in imaging as an outcome predictor and potential tool in radiotherapy treatment planning. , 2014, The British journal of radiology.

[5]  Vasilis Ntziachristos,et al.  Motion correction in optoacoustic mesoscopy , 2017, Scientific Reports.

[6]  Muhammad Imran Razzak,et al.  Deep Learning for Medical Image Processing: Overview, Challenges and Future , 2017, ArXiv.

[7]  V. Ntziachristos,et al.  Assessing hyperthermia‐induced vasodilation in human skin in vivo using optoacoustic mesoscopy , 2018, Journal of biophotonics.

[8]  Ioannis B. Theocharis,et al.  SVM-FuzCoC: A novel SVM-based feature selection method using a fuzzy complementary criterion , 2010, Pattern Recognit..

[9]  V. Ntziachristos,et al.  Assessing nailfold microvascular structure with ultra-wideband raster-scan optoacoustic mesoscopy , 2018, Photoacoustics.

[10]  Shabana Urooj,et al.  Wavelets: biomedical applications , 2015 .

[11]  Robert M. Haralick,et al.  Textural Features for Image Classification , 1973, IEEE Trans. Syst. Man Cybern..

[12]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[13]  Andrew W. Moore,et al.  Locally Weighted Learning , 1997, Artificial Intelligence Review.

[14]  G. Pellacani,et al.  Advances in non‐invasive techniques as aids to the diagnosis and monitoring of therapeutic response in plaque psoriasis: a review , 2015, International journal of dermatology.

[15]  Vasilis Ntziachristos,et al.  Implications of Ultrasound Frequency in Optoacoustic Mesoscopy of the Skin , 2015, IEEE Transactions on Medical Imaging.

[16]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[17]  David G. Stork,et al.  Pattern Classification (2nd ed.) , 1999 .

[18]  Taghi M. Khoshgoftaar,et al.  RUSBoost: A Hybrid Approach to Alleviating Class Imbalance , 2010, IEEE Transactions on Systems, Man, and Cybernetics - Part A: Systems and Humans.

[19]  J. Theocharis,et al.  Feature selection based on a fuzzy complementary criterion: application to gait recognition using ground reaction forces , 2012, Computer methods in biomechanics and biomedical engineering.

[20]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[21]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[22]  Ruikang K. Wang,et al.  In vivo volumetric imaging of microcirculation within human skin under psoriatic conditions using optical microangiography , 2011, Lasers in surgery and medicine.

[23]  Wolfgang Drexler,et al.  In situ structural and microangiographic assessment of human skin lesions with high-speed OCT , 2012, Biomedical optics express.

[24]  Sebastian Ruder,et al.  An overview of gradient descent optimization algorithms , 2016, Vestnik komp'iuternykh i informatsionnykh tekhnologii.

[25]  Vasilis Ntziachristos,et al.  Pushing the Optical Imaging Limits of Cancer with Multi-Frequency-Band Raster-Scan Optoacoustic Mesoscopy (RSOM) , 2015, Neoplasia.

[26]  M. Tanter,et al.  Ultrafast ultrasound localization microscopy for deep super-resolution vascular imaging , 2015, Nature.

[27]  June-Goo Lee,et al.  Deep Learning in Medical Imaging: General Overview , 2017, Korean journal of radiology.

[28]  V. Ntziachristos,et al.  Raster-Scanning Optoacoustic Mesoscopy for Gastrointestinal Imaging at High Resolution. , 2018, Gastroenterology.

[29]  Boris Hermann,et al.  In vivo dual-modality photoacoustic and optical coherence tomography imaging of human dermatological pathologies. , 2015, Biomedical optics express.

[30]  Chih-Kuang Yeh,et al.  Classification of scattering media within benign and malignant breast tumors based on ultrasound texture-feature-based and Nakagami-parameter images. , 2011, Medical physics.

[31]  Peter N. Heller,et al.  The application of multiwavelet filterbanks to image processing , 1999, IEEE Trans. Image Process..

[32]  Ian Witten,et al.  Data Mining , 2000 .

[33]  Wolfram Sterry,et al.  Confocal laser-scanning microscopy of capillaries in normal and psoriatic skin , 2012, Journal of biomedical optics.

[34]  Vasilis Ntziachristos,et al.  Ultrawideband reflection-mode optoacoustic mesoscopy. , 2014, Optics letters.

[35]  William A. Belson,et al.  Matching and Prediction on the Principle of Biological Classification , 1959 .