Malignant melanoma is the deadliest type of skin
cancer. If melanoma detection and diagnosis is performed in its
early stages, the probabilities of recovery and survival are higher.
Dermoscopy is a manual method which is applied by doctors to
diagnose this disease, but it strongly depends on the experience of
the specialist who performs this skin assessment. Although, many
proposals have been made for automated detection and diagnosis
of malignant melanoma based on images processing, there are
still improvement opportunities for melanoma diagnosis. This
paper aims to identify the current status of the latest researches
related to techniques for malignant melanoma diagnosis based on
images analysis, considering the three research questions that
have been elaborated for the systematic literature review: Q1)
Which are the latest methods for malignant melanoma detection?
Q2) Which systems for malignant melanoma diagnosis have been
implemented in the last 5 years? And Q3) Which CAD systems for
malignant melanoma detection have been developed?
Furthermore, a cross-analysis of the outcome was performed. The
results propose the implementation of systems using Inception V3
and the classifier Support Vector Machine, which achieved high
accuracies in malignant melanoma diagnosis based on images
processing.
hybSVM: Bacterial colony optimization algorithm based SVM for malignant melanoma detection
