Comparison of Machine Learning Algorithms Used for Skin Cancer Diagnosis

Abstract

The paper presents a comparison of automatic skin cancer diagnosis algorithms based on analyses of skin lesions photos. Two approaches are presented: the first one is based on the extraction of features from images using simple feature descriptors, and then the use of selected machine learning algorithms for the purpose of classification, and the second approach uses selected algorithms belonging to the subgroup of machine learning—deep learning, i.e., convolutional neural networks (CNN), which perform both the feature extraction and classification in one algorithm. The following algorithms were analyzed and compared: Logistic Regression, k-Nearest Neighbors, Naive Bayes, Decision Tree, Random Forest, and Support Vector Machine, and four CNN–VGG-16, ResNet60, InceptionV3, and Inception-ResNetV2 In the first variant, before the classification process, the image features were extracted using 4 different feature descriptors and combined in various combinations in order to obtain the most accurate image features vector, and thus the highest classification accuracy. The presented approaches have been validated using the image dataset from the ISIC database, which includes data from two categories—benign and malignant skin lesions. Common machine learning metrics and saved values of training time were used to evaluate the effectiveness and the performance (computational complexity) of the algorithms.

Two-step hierarchical binary classification of cancerous skin lesions using transfer learning and the random forest algorithm

“…However, the question posed in this paper pertains to the correlation between overfitting and parameter count. In such instances, opting for more traditional classifiers like support vector machines (SVMs) [ 30 ], k-nearest neighbors (k-NNs) [ 31 ], random forest (RF) [ 32 ], and logistic regression (LR) [ 32 ] becomes viable, as they require fewer parameters.…”

Skin Cancer Detection and Classification Using Neural Network Algorithms: A Systematic Review

“…They include metadata related to clinical diagnosis, lesion type, and body location, among other factors. Their open access and the diversity of data they offer make them highly valuable to the scientific community, in dermatological research, and in the development of artificial intelligence tools for the diagnosis of skin diseases as a significant complement to expert diagnosis [87][88][89][90][91]. Additionally, the PH2 dataset consists of a recompilation of 200 images, focusing on a local objective rather than being broadly applicable to other case studies [69,74].…”