A method for differentiation of pigmented skin lesions based on digital processing of optical images has been proposed. The method implies segmentation and automatic differentiation of lesions based on 5 parameters –the diameter, area, color, shape and smoothness of margins of the lesion. It reveals clinical features of the lesion, required for diagnosis, and allows for calculating a probability of the lesions malignizaion. The proposed method has been verified using 360 imaging of pigmented lesions of the skin in vivo; among them: ordinary, dysplastic nevi of the skin and skin melanoma. The observed sensitivity and specificity of the proposed technique are 97% and 95%, correspondingly.
REFERENCES
1.
N.
Eisemann
, A.
Waldmann
et al, J. Invest. Dermatol.
134
, 43
–50
(2014
).2.
S. T.
Chen
et al, Curr. Dermatol.
2
, 24
–34
(2013
).3.
V.
Nikolaou
and A. J.
Stratigos
, British Journal of Dermatology
170
, 11
–19
(2014
).4.
N.
MacKinnon
et al, Proc. of SPIE
9711
, 971111
–971117
(2016
).5.
D.
Comaniciu
and P.
Meer
, IEEE Transactions on Pattern Analysis and Machine Intelligence
24
, 603
–619
(2002
).6.
C. A.
Hussain
et al, Procedia Computer Science
85
, 571
–580
(2016
).7.
8.
S.
Mohammadi
, H. E.
Möller
et al, Magnetic Resonance in Medicine
64
, 1047
–1056
(2010
).9.
K. G.
Kudrin
, O. V.
Matorin
and I. V.
Reshetov
, Journal of Physics: Conference Series
584
, 012001
(2015
).10.
D. S.
Rigel
, J.
Russak
and R.
Friedman
, CA: A Cancer Journal for Clinician
60
, 301
–316
(2010
).
This content is only available via PDF.
© 2019 Author(s).
2019
Author(s)
You do not currently have access to this content.
