Characterization and Classification of Imbalanced Dermoscopic Datasets
DOI:
https://doi.org/10.59681/2175-4411.v15.iEspecial.2023.1085Keywords:
Skin Neoplasms, Medical Informatics, Artificial IntelligenceAbstract
Objective: To investigate computational intelligence techniques to characterize and classify imbalanced datasets of dermoscopic lesions. Methods: The developed method includes techniques for image pre-processing, feature extraction, oversampling, feature selection, and classifier building and evaluation. We assessed 20 method configurations in 274 public dermoscopies with 48 melanomas and 226 nevi. Results: We reached the highest average accuracy, 83.57%, after reducing the feature number by at least 48.86%. In general, the oversampling technique improved the average sensitivity. Conclusion: The best method results in the characterization and classification of an imbalanced dermoscopic dataset were promising and competitive with some recent references.
References
Bansal P, Garg R, Soni P. Detection of melanoma in dermoscopic images by integrating features extracted using handcrafted and deep learning models. Comput Ind Eng. 2022;168:108060.
Kaur R, GholamHosseini H, Sinha R. Hairlines removal and low contrast enhancement of melanoma skin images using convolutional neural network with aggregation of contextual information. Biomed Signal Process Control. 2022;76:103653.
Pathan S, Ali T, Vincent S, Nanjappa Y, David RM, Kumar OP. A dermoscopic inspired system for localization and malignancy classification of melanocytic lesions. Appl Sci (Basel). 2022;12(9):4243.
Popecki P, Jurczyszyn K, Ziętek M, Kozakiewicz M. Texture analysis in diagnosing skin pigmented lesions in normal and polarized light - a preliminary report. J Clin Med. 2022 Apr 29;11(9):2505.
Alazzam MB, Alassery F, Almulihi A. Diagnosis of melanoma using deep learning. Math Probl Eng. 2021;2021:1423605.
Javaid A, Sadiq M, Akram F. Skin cancer classification using image processing and machine learning. In: Zafar-Uz-Zaman M, Siddiqui NA, Iqbal M, et al., editors. Proceedings of the 18th International Bhurban Conference on Applied Sciences and Technologies; 2021; Islamabad, Pakistan. [New York]: Curran Associates; 2021. p. 439-44.
Valdés-Morales KL, Peralta-Pedrero ML, Cruz FJ, Morales-Sánchez MA. Diagnostic accuracy of dermoscopy of actinic keratosis: a systematic review. Dermatol Pract Concept. 2020;10(4):e2020121.
Lee HD, Mendes AI, Spolaôr N, Oliva JT, Sabino Parmezan AR, Chung WF, et al. Dermoscopic assisted diagnosis in melanoma: Reviewing results, optimizing methodologies and quantifying empirical guidelines. Knowl-Based Syst. 2018;158:9-24.
Instituto Nacional de Câncer (BR). Estimativa 2020: incidência de câncer no Brasil [Internet]. Rio de Janeiro: Instituto Nacional de Câncer; 2019 [cited 2022 Jul 13]. Available from: https://www.inca.gov.br/sites/ufu.sti.inca.local/files//media/document//estimativa-2020-incidencia-de-cancer-no-brasil.pdf.
Chollet F, Allaire JJ. Deep learning in R. Shelter Island: Manning publications; 2018. 335 p.
Witten IH, Frank E, Hall MA, Pal CJ. Data mining: practical machine learning tools and techniques. 4a. ed. Burlington: Morgan Kaufmann; 2016. 654 p.
Liu H, Motoda H. Computational methods of feature selection. Boca Ratón: Chapman & Hall/CRC; 2007. 411 p.
Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP. SMOTE: synthetic minority over-sampling technique. J Artif Intell Res. 2002;16(1):321-57.
Grezzana APB, Lee HD, Spolaôr N, Wu FC. Extração e Seleção de Atributos para Processamento e Análise de Imagens Médicas. In: Pró-Reitoria de Pesquisa da Universidade de São Paulo, editor. Anais do Simpósio Internacional de Iniciação Científica e Tecnológica da USP; 2021; São Carlos, Brasil. São Paulo: Universidade de São Paulo; 2021. p. 1-1.
Grezzana APB, Lee HD, Spolaôr N, Wu FC. Segmentação, Caracterização e Classificação de Imagens Dermoscópicas Usando Seleção de Atributos. In: Comitê Assessor de Bolsas de Iniciação Científica da Universidade Estadual do Oeste do Paraná, editor. Anais do Encontro Anual de Iniciação Científica, Tecnológica e Inovação da Unioeste; 2021; Cascavel, Brasil. Cascavel: Universidade Estadual do Oeste do Paraná; 2021. p. 1-1.
Merck Sharp & Dohme. Nevos [Internet]. Rahway: Merck Sharp & Dohme; 2020 [cited 2022 Jul 13]. Available from: https://www.msdmanuals.com/pt-br/casa/dist%C3%BArbios-da-pele/tumores-cut%C3%A2neos-n%C3%A3o-cancerosos/nevos.
Kuhn M, Johnson K. Applied predictive modelling. New York: Springer; 2013. 613 p.
Haralick RM, Shanmugam K, Dinstein I. Textural features for image classification. IEEE Trans Syst Man Cybern. 1973;SMC-3(6):610-21.
Laws KI. Texture energy measures. In: Baumann LS, editor. Proceedings of the Defense Advanced Research Projects Agency Image Understanding Workshop; 1979; Los Angeles, United States. [Arlington]: Science Applications; [1979?]. p. 47-51.
Smit S, Hoefsloot HCJ, Smilde AK. Statistical data processing in clinical proteomics. J Chromatogr B Analyt Technol Biomed Life Sci. 2008;866(1):77-88.
Carvalho VAM, Spolaôr N, Cherman EA, Monard MC. A Framework for Multi-label Exploratory Data Analysis: ML-EDA. In: Ezzatti P, Delgado A, editors. Proceedings of the Latin American Computing Conference; 2014; Montevidéu, Uruguai. [New York]: Curran Associates; 2014. p. 1-12.
Oliva JT, Lee HD, Spolaôr N, Coy CSR, Chung WF. Prototype system for feature extraction, classification and study of medical images. Expert Syst and Appl. 2016;63:267-83.
Amadasun M, King R. Textural features corresponding to textural properties. IEEE Trans Syst Man Cybern. 1989;19(5):1264-74.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2023 Newton Spolaôr, Huei Diana Lee, Weber Shoity Resende Takaki, Leandro Augusto Ensina, Antonio Rafael Sabino Parmezan, Matheus Maciel, Claudio Saddy Rodrigues Coy, Feng Chung Wu
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Submission of a paper to Journal of Health Informatics is understood to imply that it is not being considered for publication elsewhere and that the author(s) permission to publish his/her (their) article(s) in this Journal implies the exclusive authorization of the publishers to deal with all issues concerning the copyright therein. Upon the submission of an article, authors will be asked to sign a Copyright Notice. Acceptance of the agreement will ensure the widest possible dissemination of information. An e-mail will be sent to the corresponding author confirming receipt of the manuscript and acceptance of the agreement.
