Computer-Supported Diagnosis of Melanoma in Profilometry

Abstract:

Laser profilometry offers new possibilities to improve non-invasive tumor diagnostics in dermatology. In this paper, a new approach to computer-supported analysis and interpretation of high-resolution skin-surface profiles of melanomas and nevocellular nevi is presented. Image analysis methods are used to describe the profile‘s structures by texture parameters based on co-occurrence matrices, features extracted from the Fourier power spectrum, and fractal features. Different feature selection strategies, including genetic algorithms, are applied to determine the best possible subsets of features for the classification task. Several architectures of multilayer perceptrons with error back-propagation as learning paradigm are trained for the automatic recognition of melanomas and nevi. Furthermore, network-pruning algorithms are applied to optimize the network topology. In the study, the best neural classifier showed an error rate of 4.5% and was obtained after network pruning. The smallest error rate in all, of 2.3%, was achieved with nearest neighbor classification.

Keywords:

Laser Profilometry - Image Analysis - Feature Selection - Pattern Recognition - Artificial Neural Networks

• REFERENCES

• 1 Elwood JM, Koh HK. Etiology, Epidemiology, Risk Factors and Public Health Issues of Melanoma. Current Opinions in Oncology 1994; 6: 179.
• 2 Balch CM, Milton GW. Hautmelanome. Berlin: Springer Verlag; 1988
• 3 Grin CM, Kopf AW, Welkovich B. Accuracy in the Clinical Diagnosis of Malignant Melanoma. Arch Dermatol 1990; 126: 763.
• 4 Golston JE, Stoecker WV, Moss RH, Dhillon IPS. Automatic Detection of Irregular Borders in Melanoma and Other Skin Tumors. Computerized Medical Imaging and Graphics 1992; 16 (Suppl. 03) 163-77.
• 5 Green A, Martin N, Pfitzner J, O’Rourke M, Knight N. Computer Image Analysis in the Diagnosis of Melanoma. J Am Acad Dermatol 1994; 31 (Suppl. 06) 958-64.
• 6 Sober AJ, Burstein JM. Computerized Digital Image Analysis: An Aid for Melanoma DiagnosisPreliminary Investigations and Brief Review. Journal of Dermatology 1994; 2 (Suppl. 11) 885-90.
• 7 Wilhelm KP, Elsner P, Beradesca E, Mai-bach H. (eds). Bioengineering of the Skin: Skin Surface Imaging and Analysis. Boca Raton: CRC Press; 1997
• 8 Handels H, Roß T, Kreusch J, Wolf HH, Pöppl SJ. Image Analysis and Pattern Recognition to Support Skin Tumor Diagnosis. In: Cesnik B, McCray AT, Scherrer JR. (eds). MEDINFO ’98. Amsterdam: IOS Press; 1998: 1056-62.
• 9 Handels H. Mustererkennung, Bildanalyse und Visualisierung für die computerunterstützte, ärztliche Diagnostik – Methoden und Anwendungen zur Analyse und Erkennung von Tumoren in medizinischen Bilddaten, Habilitationsschrift. Lübeck: Medizinische Universität zu Lübeck; 1998
• 10 Roß T. Analyse und Klassifikation zweidimensionaler Signale – Anwendung in der Erkennung von Hauttumoren anhand hochaufgelöster Oberflächenprofile, Ph.D. thesis, Medical University of Lübeck, Aachen: Shaker Verlag; 1997
• 11 Kreusch J, Busche H, Connemann BJ. et al. Differentiation between Nevocellular Nevi and Malignant Melanoma by Skin Surface Parameters. In: Wilhelm KP, Elsner P, Beradesca E, Maibach H. eds. Bioengineering of the Skin: Skin Surface Imaging and Analysis, Boca Raton: CRC Press; 1997: 289-300.
• 12 Bondi EE, Elder DE, Dupont G, Clark WH. Skin Markings in Malignant Melanoma. J Am Med Assoc. 1984; 250: 503.
• 13 Busche H. Vergleichende Untersuchung der quantitativen Oberflächentopographie superfiziell spreitzender Melanome unnd nävozellulärer Nävi, master thesis,. Medical University of Lübeck; 1994
• 14 Foley J, Dam vA SF, Hughes J. Computer Graphics: Principles and Practice. Addison Wesley; 1990
• 15 Haralick RM, Shanmugam K, Dinstein I. Textural Features for Image Classification. IEEE Transaction on Systems Man Cybernetics 1973; 3: 610-21.
• 16 Ballard DH, Brown MB. Computer Vision. Englewood Cliffs: PrenticeHall; 1982
• 17 Gotlieb CC, Kreyszig HE. Texture Descriptors Based on Co-occurrence Matrices. Computer Vision, Graphics and Image Processing 1990; 51: 70-86.
• 18 Falconer KJ. Fractal Geometry. Mathematical Foundations and Applications, Chi-chester: John Wiley & Sons; 1990
• 19 Peitgen HO, Saupe D. The Science of Fractal Images. Berlin: Springer Verlag; 1988
• 20 Campenhout JMv. On the Peaking of the Hughes Mean Recognition Accuracy – the Resolution of an Apparent Paradox. IEEE Transactions on Systems, Man and Cybernetics 1978; SMC-8: 390-5.
• 21 Jain AK, Waller WG. On the Optimal Number of Features in the Classification of Multivariate Gaussian Data. Pattern Recognition 1978; 10: 365-74.
• 22 Jain AK. Advances in Statistical Pattern Recognition. In: Devijer PA, Kittler J. eds. Pattern Recognition, Theory and Applications. Berlin: Springer Verlag; 1986
• 23 Jelonek J, Stefanowski J. Feature Subset Selection for Classification of Histological Images. Artificial Intelligence in Medicine 1997; 9 (Suppl. 03) 227-40.
• 24 Kanal L, Chandrasekaran B. On Dimensionality and Sample Size in Statistical Pattern Recognition. Pattern Recognition 1971; 3: 225-34.
• 25 Kohavi R, SommerfIeld D. Feature Subset Selection using the Wrapper method: Over-fitting and Dynamic Search Space Topology. 1st Int. Conf. Knowledge Discovery Data Mining. Montreal: AAAI Press; 1995: 192-7.
• 26 Niemann H. Klassifikation von Mustern. Berlin: Springer Verlag; 1983
• 27 Raudys S, Jain AK. Small Sample Size Problems in Designing Artificial Nerual Networks. In: Sethi IK, Jain AK. eds. Artificial Neural Networks and Statistical Pattern Recognition. Amsterdam: NorthHolland; 1991: 33-50.
• 28 Siedlecki W, Sklansky J. A Note on Genetic Algorithms for Large scale Feature Selection. Pattern Recognition Letters 1989; 10: 335-47.
• 29 Handels H, Roß T, Kreusch J, Wolf HH, Pöppl SJ. A New Approach for Automatic Recognition of Melanoma in Profilometry: Optimized Feature Selection using Genetic Algorithms. In: Hanson KM. (ed). SPIE Medical Imaging 1998. San Diego: SPIE, Vol. 3338; 1998: 684-92.
• 30 Davis L. Handbook of Genetic Algorithms. New York: Van Nostrand Reinhold; 1991
• 31 Goldberg DE. Genetic Algorithms in Search. Optimization and Machine Learning, Massachusetts: Addison Wesley; 1989
• 32 Holland JH. Adaption in Natural and Artificial Systems. Michigan: University of Michigan Press; 1975
• 33 Rumelhart DE, Hinton GE, Williams RJ. Learning Representations by Back-Propagation Errors. Nature 1986; 323: 533-6.
• 34 Rumelhart DE, Hinton GE, Williams RJ. Learning Internal Representations by Error Propagation. In: Rumelhart DE, McClelland JL. eds. Parallel Distributed Processing. Exploration in the Microstructures of Cognition. Cambridge: MIT Press; 1986
• 35 Carpenter GA, Grossberg S. Neural Networks for Vision and Image Processing. London: A Bradford Book, MIT Press; 1992
• 36 Miller AS, Blott BH, Harmes TK. Review of Neural Network Applications in Medical Imaging and Signal Processing. Medical & Biological Engineering & Computing 1992; 30: 449-64.
• 37 Schürmann J. Pattern Classification. New York: John Wiley & Sons; 1996
• 38 Reed R. Pruning Algorithms A Survey. IEEE Transactions on Neural Networks 1993; 45: 740-7.