AN INNOVATIVE TECHNIQUE FOR CLASSIFICATION AND DETECTION OF SKIN CANCER USING IMAGE PROCESSING AND SVM ALGORITHM

Abstract

Skin cancer is a serious health issue that affects millions of people worldwide. One in every three cancers is a skin cancer and, most people fail to identify and get a diagnosis which is why early detection is essential for effective treatment and improving patient outcomes. In recent years, computer vision and machine learning have become important tools for the automatic detection of skin cancer. One of the commonly used machine learning algorithms for this task is Support Vector Machines (SVMs). SVMs are a type of supervised learning algorithm that is used for classification tasks. In skin cancer detection, the SVM classifier is trained on a dataset of dermatoscopic images of skin lesions. The first step in the process is feature extraction, where relevant information is extracted from the images to serve as input to the SVM classifier. This information can include color, texture, and shape features, among others.
 The training and testing of the SVM classifier is then performed using a portion of the dataset, with the remainder being used to evaluate its performance. During the testing phase, the SVM classifier is used to predict the class label of each image, which can be malignant or benign. The accuracy of the classifier is evaluated by comparing its predictions to the actual class labels of the images in the evaluation dataset.
 The results of using SVMs in skin cancer detection have been promising, with high accuracy rates being achieved. This highlights the potential of SVMs as a useful tool for skin cancer screening and early detection. In conclusion, the use of SVMs in skin cancer detection provides a fast, automatic, and reliable method for detecting skin cancer, which can help to improve patient outcomes. Currently it is really very important to watch and analyse the cancer disease automatically at intervals the first stages. Irregular streaks square measure one in every of the foremost very important features (included in most of dermoscopy algorithms) that show high association with carcinoma and basal cell malignant growth malady. The diagnostic test technique for the detection is most painful and harmful so we have a tendency to tend to square measure going for the machine-driven detection. Here we have a tendency to tend to square measure practice the GLCM choices for the detection the choices of skin lesions square measure extracted normalized symmetrical grey Level Co-occurrence Matrices GLCM. GLCM based texture choices square measure extracted from each of the four classes and given as input to the Multi-Class Support vector machine that's utilized for classification purpose.