Olusola Oluwakemi ABAYOMI-ALLI, Robertas DAMAŠEVIČIUS, Sanjay MISRA, Rytis MASKELIŪNAS, Adebayo ABAYOMI-ALLI

Malignant skin melanoma detection using image augmentation by oversampling in nonlinear lower-dimensional embedding manifold

The continuous rise in skin cancer cases, especially in malignant melanoma, has resulted in a high mortality rate of the affected patients due to late detection. Some challenges affecting the success of skin cancer detection include small datasets or data scarcity problem, noisy data, imbalanced data, inconsistency in image sizes and resolutions, unavailability of data, reliability of labeled data (ground truth), and imbalance of skin cancer datasets. This study presents a novel data augmentation technique based on covariant Synthetic Minority Oversampling Technique (SMOTE) to address the data scarcity and class imbalance problem. We propose an improved data augmentation model for effective detection of melanoma skin cancer. Our method is based on data oversampling in a nonlinear lower-dimensional embedding manifold for creating synthetic melanoma images. The proposed data augmentation technique is used to generate a new skin melanoma dataset using dermoscopic images from the publicly available P H2 dataset. The augmented images were used to train the SqueezeNet deep learning model. The experimental results in binary classification scenario show a significant improvement in detection of melanoma with respect to accuracy (92.18%), sensitivity (80.77%), specificity (95.1%), and F1-score (80.84%). We also improved the multiclass classification results in melanoma detection to 89.2% (sensitivity), 96.2% (specificity) for atypical nevus detection, 65.4% (sensitivity), 72.2% (specificity), and for common nevus detection 66% (sensitivity), 77.2% (specificity). The proposed classification framework outperforms some of the state-of-the-art methods in detecting skin melanoma.

PDF

___

[1] Codella NC, Nguyen QB, Pankanti S, Gutman DA, Helba B et al. Deep learning ensembles for melanoma recognition in dermoscopy images. IBM Journal of Research and Development. 2017; 61 (4/5):5–1.
[2] Ahmed H, Al-azawi R, Abdulhameed A. Evaluation methodology between globalization and localization features approaches for skin cancer lesions classification. In: Journal of Physics: Conference Series. vol. 1003; 2018. p. 012029.
[3] Barata C, Ruela M, Francisco M, Mendonca T, Marques JS. Two systems for the detection of melanomas in dermoscopy images using texture and color features. IEEE Systems Journal. 2013; 8 (3):965–979.
[4] Brinker TJ, Hekler A, Enk AH, Berking C, Haferkamp S et al. Deep neural networks are superior to dermatologists in melanoma image classification. European Journal of Cancer. 2019;119:11–17.
[5] Ratul MAR, Mozaffari MH, Lee W, Parimbelli E. Skin Lesions Classification Using Deep Learning Based on Dilated Convolution. bioRxiv. 2020:860700.
[6] Oh BH, Kim KH, Chung KY. Skin Imaging Using Ultrasound Imaging, Optical Coherence Tomography, Confocal Microscopy, and Two-Photon Microscopy in Cutaneous Oncology. Frontiers in Medicine. 2019;6:274.
[7] Barata C, Celebi ME, Marques JS. A survey of feature extraction in dermoscopy image analysis of skin cancer.IEEE journal of biomedical and health informatics. 2018; 23 (3):1096–1109.
[8] Li H, He X, Zhou F, Yu Z, Ni D et al. Dense deconvolutional network for skin lesion segmentation. IEEE Journal of Biomedical and Health Informatics. 2018; 23 (2):527–537.
[9] Maiti A, Chatterjee B. Improving detection of Melanoma and Naevus with deep neural networks. Multimedia Tools and Applications. 2019:1–20.
[10] Zhang J, Xie Y, Xia Y, Shen C. Attention residual learning for skin lesion classification.IEEE transactions on medical imaging. 2019; 38 (9):2092–2103.
[11] Hosny KM, Kassem MA, Foaud MM. Classification of skin lesions using transfer learning and augmentation with Alex-net. PloS one. 2019; 14 (5):e0217293.
[12] Pacheco AG, Krohling RA. The impact of patient clinical information on automated skin cancer detection. Computers in biology and medicine. 2020;116:103545
[13] Curiel-Lewandrowski C, Berry E, Leachman S. Artificial intelligence approach in melanoma. Melanoma New York (NY): Springer. 2019:1–31.
[14] Lateh MA, Muda AK, Yusof ZIM, Muda NA, Azmi MS. Handling a Small Dataset Problem in Prediction Model by employ Artificial Data Generation Approach: A Review. Journal of Physics: Conference Series. 2017 Sep;892:012016.
[15] Kassani SH, Kassani PH. A comparative study of deep learning architectures on melanoma detection. Tissue and Cell. 2019; 58:76–83.
[16] Nasr-Esfahani E,Rafiei S,Jafari MH,Karimi N, Wrobel JS et al. Dense pooling layers in fully convolutional network for skin lesion segmentation.Computerized Medical Imaging and Graphics. 2019;78:101658.
[17] Hekler A,Utikal JS,Enk AH,Berking C, Klode J et al. Pathologist-level classification of histopathological melanoma images with deep neural networks. European Journal of Cancer. 2019;115:79–83.
[18] Zhang C, Wang X, Zhang F, Dong L, Wu J, Cheng Q, et al. Phenylalanine ammonia-lyase contributes to the soybean response towards Phytophthora sojae infection. Scientific reports. 2017; 7 (1):1–13.
[19] Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP. SMOTE: synthetic minority over-sampling technique. Journal of artificial intelligence research. 2002;16:321–357.
[20] Khan MA, Akram T, Sharif M, Javed K, Rashid M et al. An integrated framework of skin lesion detection and recognition through saliency method and optimal deep neural network features selection. Neural Computing and Applications. 2019:1–20.
[21] Abayomi-Alli OO, Damaševičius R, Wieczorek M, Wozniak M. Data Augmentation Using Principal Component Resampling for Image Recognition by Deep Learning. In: International Conference on Artificial Intelligence and Soft Computing. Springer; 2020. pp. 39–48.
[22] Urbonas A, Raudonis V, Maskeliunas R, Damasevicius R. Automated identification of wood veneer surface defects using faster region-based convolutional neural network with data augmentation and transfer learning. Applied Sciences. 2019; 9 (22):4898.
[23] Leguen-deVarona I, Madera J, Martinez-Lopez Y, Hernandez-Nieto JC. SMOTE-Cov: A New Oversampling Method Based on the Covariance Matrix. In: Data Analysis and Optimization for Engineering and Computing Problems.Springer; 2020. pp. 207–215.
[24] Iandola FN, Han S, Moskewicz MW, Ashraf K, Dally WJ et al. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size; 2016.
[25] Munir K, Elahi H, Ayub A, Frezza F, Rizzi A.Cancer diagnosis using deep learning: a bibliographic review. Cancers. 2019; 11 (9):1235.
[26] Roda A, Michelini E, Zangheri M, Di Fusco M, Calabria D et al. Smartphone-based biosensors: A critical review and perspectives. TrAC Trends in Analytical Chemistry. 2016;79:317–325.
[27] Rahimzadeganasl A, Alganci U, Goksel C. An Approach for the Pan Sharpening of Very High Resolution Satellite Images Using a CIELab Color Based Component Substitution Algorithm. Applied Sciences. 2019; 9 (23):5234.
[28] Qiu J, Harold Li H, Zhang T, Ma F, Yang D. Automatic x-ray image contrast enhancement based on parameter auto-optimization. Journal of applied clinical medical physics. 2017; 18 (6):218–223.
[29] Kim SJ, Min BS, Lim DK, Lee JH. Determining parameters in contrast limited adaptive histogram equalization. In: Proceedings of the 7th International Conference on Information Security and Assurance. vol. 21; 2013. pp. 204–207.
[30] Frangi AF, Niessen WJ, Vincken KL, Viergever MA. Multiscale vessel enhancement filtering. In: International conference on medical image computing and computer-assisted intervention. Springer; 1998. pp. 130–137.
[31] Mumford DB, Shah J. Optimal approximations by piecewise smooth functions and associated variational problems. Communications on pure and applied mathematics. 1989.
[32] Sahlol AT,Abd Elaziz M, Tariq Jamal A,Damasevicius R,Farouk Hassan O. A novel method for detection of tuberculosis in chest radiographs using artificial ecosystem-based optimisation of deep neural network features. Symmetry. 2020;12 (7):1146.25
[33] Mendonca T,Ferreira PM,Marques JS,Marcal AR,Rozeira J. PH 2-A dermoscopic image database for research and benchmarking. In: 2013 35th annual international conference of the IEEE engineering in medicine and biology society (EMBC). IEEE; 2013. pp. 5437–5440.
[34] Khan MA, Javed MY, Sharif M, Saba T, Rehman A. Multi-model deep neural network based features extraction and optimal selection approach for skin lesion classification. In: 2019 international conference on computer and information sciences (ICCIS). IEEE; 2019. pp. 1–7.
[35] Hu K, Niu X, Liu S, Zhang Y, Cao C, Xiao F, et al. Classification of melanoma based on feature similarity measurement for codebook learning in the bag-of-features model. Biomedical Signal Processing and Control. 2019; 51:200–209.
[36] Adjed F, Gardezi SJS,Ababsa F,Faye I,Dass SC. Fusion of structural and textural features for melanoma recognition. IET Computer Vision. 2017; 12 (2):185–195.
[37] Eltayef K, Li Y, Liu X. Detection of melanoma skin cancer in dermoscopy images. In: Journal of Physics: Conference Series. vol. 787. IOP Publishing; 2017. p. 012–034.
[38] Bi L, Kim J, Ahn E, Feng D, Fulham M. Automatic melanoma detection via multi-scale lesion-biased representation and joint reverse classification. In: 2016 IEEE 13th international symposium on biomedical imaging (ISBI). IEEE; 402016. p. 1055–1058.