Melanoma, often known as malignant melanoma, is the commonest yet deadliest kind of skin cancer. While melanoma can be prevented to some extent by educating, the public about safe sun activities like avoiding high radiation hours, wearing protective clothing, trying to apply sunscreen, and keeping a safe distance UV light sources that are generated artificially, early diagnosis and the exact treatment of illness can help to reduce the fatality rate. Efficient and early appropriate treatment of melanoma has been a priority for researchers and the doctors, numerous invasive and non-invasive approaches for melanoma diagnosis have come into focus from time to time. Easy access to skin exams increases the likelihood of accurate and timely identification of melanoma, according to an analysis of different approaches established over the years, and computer-assisted diagnosis (CAD) has played a vital part in achieving this goal. We have proposed a convolutional neural network with eight thick layers for the categorization of melanoma lesions in this research. Inception and Residual blocks are used to extraction of features at both local and global level. When tested on the International Skin Imaging Collaboration (ISIC) 2018, ISIC 2019, and ISIC 2020 datasets, the suggested classifier has a depth of 40 layers, allowing it to attain a high accuracy score.
CITATION STYLE
Singh, S. K., Banerjee, S., Chakraborty, A., & Bandyopadhyay, A. (2023). Classification of Melanoma Skin Cancer Using Inception-ResNet. In Lecture Notes in Networks and Systems (Vol. 519 LNNS, pp. 65–74). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-19-5191-6_6
Mendeley helps you to discover research relevant for your work.