Detection of Stage of Cancer Using Machine Learning and Deep Learning

Melanoma, an aggressive form of skin cancer known for its rapid spread, poses a significant threat and is responsible for a substantial number of fatalities. The accurate classification of cancer stages is crucial for effective diagnosis, particularly during surgical treatment. This paper introduces two methods for categorizing melanoma cancer stages. Initial system distinguishes between stage 1 and stage 2, while the second system extends the classification to include stage 3 melanoma. Leveraging a Convolutional Neural Network (CNN) algorithm, the proposed system incorporates the Similarity Measure for Text Processing (SMTP) as a loss function. The experimental results showcase the efficacy of different loss functions, with a particular focus on the proposed SMTP loss function, which outperforms several others designed for classification problems. This algorithm stands out as a more efficient approach for melanoma stage classification.