Lentigo detection using a deep learning approach

Reflectance confocal microscopy (RCM) allows fast data acquisition with a high resolution of the skin. In fact, RCM images are becoming more and more used for lentigo diagnosis. In this paper, we propose a new classification method to automate specific steps in lentigo diagnosis. Our method is based on a convolutional neural network (CNN) on InceptionV3 architecture combined with data augmentation and transfer learning. The experimental validation showed the efficiency of our model by reaching an accuracy of 98.14%.

[1]  L A Bollea-Garlatti,et al.  Lentigo Maligna: Keys to Dermoscopic Diagnosis. , 2016, Actas dermo-sifiliograficas.

[2]  Peter König,et al.  Further Advantages of Data Augmentation on Convolutional Neural Networks , 2018, ICANN.

[3]  Bram van Ginneken,et al.  A survey on deep learning in medical image analysis , 2017, Medical Image Anal..

[4]  Giovanni Pellacani,et al.  Reflectance Confocal Microscopy for In Vivo Skin Imaging † , 2008, Photochemistry and photobiology.

[5]  Mauricio Marengoni,et al.  A Survey of Transfer Learning for Convolutional Neural Networks , 2019, 2019 32nd SIBGRAPI Conference on Graphics, Patterns and Images Tutorials (SIBGRAPI-T).

[6]  Yoshua Bengio,et al.  How transferable are features in deep neural networks? , 2014, NIPS.

[7]  Taghi M. Khoshgoftaar,et al.  A survey on Image Data Augmentation for Deep Learning , 2019, Journal of Big Data.

[8]  Jean-Yves Tourneret,et al.  Statistical modeling and classification of reflectance confocal microscopy images , 2017, 2017 IEEE 7th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP).

[9]  Richard K. G. Do,et al.  Convolutional neural networks: an overview and application in radiology , 2018, Insights into Imaging.

[10]  Jean-Yves Tourneret,et al.  An unsupervised Bayesian approach for the joint reconstruction and classification of cutaneous reflectance confocal microscopy images , 2017, 2017 25th European Signal Processing Conference (EUSIPCO).

[11]  K J Dallimore,et al.  Lentigo maligna and lentigo maligna melanoma. , 1997, Journal of the American Academy of Dermatology.

[12]  Eve Oualid.  Les produits d epigmentants : le point en 2011 , 2011 .

[13]  R. Webb,et al.  In vivo confocal scanning laser microscopy of human skin: melanin provides strong contrast. , 1995, The Journal of investigative dermatology.

[14]  Sergey Ioffe,et al.  Rethinking the Inception Architecture for Computer Vision , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Y. Benezeth,et al.  Two Schemes for Automated Diagnosis of Lentigo on Confocal Microscopy Images , 2019, 2019 IEEE 4th International Conference on Signal and Image Processing (ICSIP).

[16]  R. Labadie,et al.  Histopathology of solar lentigines of the face: a quantitative study. , 1997, Journal of the American Academy of Dermatology.