Abstract
Now a days, skin cancer is the most common disease of death amongst human community. Skin cancer is exponential abnormal growth of skin cells that develops on the entire body exposed to the sunlight. Most of the skin cancers are remediable at beginning stages. In this case, the early analysis and prediction with fast detection of skin cancer to protect the human life. In this paper, we have proposed a skin cancer identification system using different soft computing techniques which are used to detect skin cancer disease in early stages. Diagnosing methodology uses image processing techniques, several machine learning and soft computing techniques for classification which includes SVM, ANN, KNN, and ensemble classifiers. The raw image of skin cancer is taken under various pre-processing techniques which are used to noise removal, enhancement and segmentation. Some features of the image have to be extracted using Gray Level Co-occurrence Matrix (GLCM) with features given as the input to classifier. SVM, ANN, KNN, and ensemble classifiers are used for classification and given image into skin cancer or normal. Numerical illustrations were also provided to prove the results and discussions.
Keyword
Grayscale image, image filter, image enhancement, image segmentation, prediction, and machine learning.
PDF Download (click here)
|