A computer‐aided melanoma detection using deep learning and an improved African vulture optimization algorithm. Issue 6 (14th April 2022)

• Record Type: Journal Article
• Authors: Hu, Libing; Zhang, Yongchun; Chen, Kaidi; Mobayen, Saleh
• Is Part Of: International journal of imaging systems and technology. Volume 32:Issue 6(2022)
• Page Start: 2002
• Page End: 2016
• Publication Date: 2022-04-14
• Subjects: computer‐aided diagnosis -- convolutional neural network -- improved African vulture optimization algorithm -- melanoma -- SIIM‐ISIC melanoma dataset; Imaging systems -- Periodicals; Image processing -- Periodicals; 621.367
• DOI: 10.1002/ima.22738
• Languages: English
• ISSNs: 0899-9457

Abstract

Abstract: Melanoma is a type of skin cancer that is caused by the uncontrolled growth of melanocytes. Cancer begins when cells in the body begin to grow out of control. Cells in almost every part of the body can become cancerous and then spread to other parts of the body. Melanoma is much less common than other types of skin cancer, such as basal cell carcinoma and squamous cell carcinoma, but melanoma is more dangerous because it can spread to other parts of the body if left undiagnosed and untreated. Melanoma is the deadliest type of skin cancer and yearly causes 60 000 people deaths. However, if it is diagnosed at an early stage, the cure rate can increase by up to 95%. The present study proposes a new procedure for the optimal diagnosis of malignant melanoma based on a new combined convolutional neural network and a newly improved metaheuristic algorithm. In this study, after applying the preprocessing technique, which contains Kapur segmentation and mathematical morphology, the key features of the target region are extracted from the image to make simpler data for processing. Afterward, a convolutional neural network (CNN) has been employed for providing the diagnosis system. To design the optimal diagnosis, we propose a newly developed design of an African Vulture Optimizer for the optimal configuration of the CNN. We also verify the effectiveness of the suggested approach based on a popular dataset, called the SIIM‐ISIC Melanoma dataset, and a comparison of its achievements with several other approaches from the literature is carried out to indicate its effectiveness.