DermaKNet: Incorporating the Knowledge of Dermatologists to Convolutional Neural Networks for Skin Lesion Diagnosis

Traditional approaches to automatic diagnosis of skin lesions consisted of classifiers working on sets of hand-crafted features, some of which modeled lesion aspects of special importance for dermatologists. Recently, the broad adoption of convolutional neural networks (CNNs) in most computer vision tasks has brought about a great leap forward in terms of performance. Nevertheless, with this performance leap, the CNN-based computer-aided diagnosis (CAD) systems have also brought a notable reduction of the useful insights provided by hand-crafted features. This paper presents DermaKNet, a CAD system based on CNNs that incorporates specific subsystems modeling properties of skin lesions that are of special interest to dermatologists aiming to improve the interpretability of its diagnosis. Our results prove that the incorporation of these subsystems not only improves the performance, but also enhances the diagnosis by providing more interpretable outputs.

[1]  Yi Su,et al.  A Novel Multi-task Deep Learning Model for Skin Lesion Segmentation and Classification , 2017, ArXiv.

[2]  Eduardo Valle,et al.  RECOD Titans at ISIC Challenge 2017 , 2017, ArXiv.

[3]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[4]  Paul Nghiem,et al.  Interactive Atlas of Dermoscopy , 2004 .

[5]  Ali Madooei,et al.  Intrinsic Melanin and Hemoglobin Colour Components for Skin Lesion Malignancy Detection , 2012, MICCAI.

[6]  Qiang Chen,et al.  Accurate and Scalable System for Automatic Detection of Malignant Melanoma , 2015 .

[7]  K Wolff,et al.  In vivo epiluminescence microscopy of pigmented skin lesions. I. Pattern analysis of pigmented skin lesions. , 1987, Journal of the American Academy of Dermatology.

[8]  Luc Van Gool,et al.  The Pascal Visual Object Classes Challenge: A Retrospective , 2014, International Journal of Computer Vision.

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

[10]  Ross B. Girshick,et al.  Fast R-CNN , 2015, 1504.08083.

[11]  Riccardo Bono,et al.  Objective follow-up of atypical melanocytic skin lesions: a retrospective study , 2010, Archives of Dermatological Research.

[12]  Noel C. F. Codella,et al.  Skin lesion analysis toward melanoma detection: A challenge at the 2017 International symposium on biomedical imaging (ISBI), hosted by the international skin imaging collaboration (ISIC) , 2016, 2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018).

[13]  Aurora Sáez,et al.  Model-Based Classification Methods of Global Patterns in Dermoscopic Images , 2014, IEEE Transactions on Medical Imaging.

[14]  Martin A Weinstock,et al.  Cutaneous melanoma: public health approach to early detection , 2006, Dermatologic therapy.

[15]  Hao Chen,et al.  Automated Melanoma Recognition in Dermoscopy Images via Very Deep Residual Networks , 2017, IEEE Transactions on Medical Imaging.

[16]  W. Stolz,et al.  The ABCD rule of dermatoscopy. High prospective value in the diagnosis of doubtful melanocytic skin lesions. , 1994, Journal of the American Academy of Dermatology.

[17]  Fernando Díaz-de-María,et al.  Enriched dermoscopic-structure-based cad system for melanoma diagnosis , 2017, Multimedia Tools and Applications.

[18]  J. Ferlay,et al.  Cancer incidence and mortality patterns in Europe: estimates for 40 countries in 2012. , 2013, European journal of cancer.

[19]  Dr. Kailash Shaw,et al.  Skin Lesion Analysis towards Melanoma Detection , 2018 .

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

[21]  Sule Yildirim Yayilgan,et al.  Combining deep learning and hand-crafted features for skin lesion classification , 2016, 2016 Sixth International Conference on Image Processing Theory, Tools and Applications (IPTA).

[22]  David Dagan Feng,et al.  Automatic Skin Lesion Analysis using Large-scale Dermoscopy Images and Deep Residual Networks , 2017, ArXiv.

[23]  Hiroshi Koga,et al.  Image Classification of Melanoma, Nevus and Seborrheic Keratosis by Deep Neural Network Ensemble , 2017, ArXiv.

[24]  Begoña Acha,et al.  Pattern analysis of dermoscopic images based on Markov random fields , 2009, Pattern Recognit..

[25]  David I. McLean,et al.  Global pattern analysis and classification of dermoscopic images using textons , 2012, Medical Imaging.

[26]  Antonio Pietrosanto,et al.  Automatic Diagnosis of Melanoma Based on the 7-Point Checklist , 2014 .

[27]  Pavel Kisilev,et al.  Medical Image Description Using Multi-task-loss CNN , 2016, LABELS/DLMIA@MICCAI.

[28]  Andrew Zisserman,et al.  A Statistical Approach to Texture Classification from Single Images , 2004, International Journal of Computer Vision.

[29]  Mohammad Toossi,et al.  Early Detection of Melanoma in Dermoscopy of Skin Lesion Images by Computer Vision–Based System , 2015 .

[30]  Yoshua Bengio,et al.  Understanding the difficulty of training deep feedforward neural networks , 2010, AISTATS.

[31]  Sebastian Thrun,et al.  Dermatologist-level classification of skin cancer with deep neural networks , 2017, Nature.

[32]  Trevor Darrell,et al.  Fully Convolutional Networks for Semantic Segmentation , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[33]  Begoña García Zapirain,et al.  Melanomas non-invasive diagnosis application based on the ABCD rule and pattern recognition image processing algorithms , 2011, Comput. Biol. Medicine.

[34]  Iván González-Díaz,et al.  Incorporating the Knowledge of Dermatologists to Convolutional Neural Networks for the Diagnosis of Skin Lesions , 2017, ArXiv.

[35]  Trevor Darrell,et al.  Constrained Convolutional Neural Networks for Weakly Supervised Segmentation , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[36]  Stein Olav Skrøvseth,et al.  Performance of a dermoscopy-based computer vision system for the diagnosis of pigmented skin lesions compared with visual evaluation by experienced dermatologists , 2014, Artif. Intell. Medicine.

[37]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[38]  A. Paolillo,et al.  A software tool for the diagnosis of melanomas , 2010, 2010 IEEE Instrumentation & Measurement Technology Conference Proceedings.

[39]  John R. Smith,et al.  Deep Learning, Sparse Coding, and SVM for Melanoma Recognition in Dermoscopy Images , 2015, MLMI.