Abstract: Digital Image Processing (DIP) plays an important role in the process of segmentation and classification of biomedical images. The detection of melanoma from skin images is the most widely used in biomedical imaging applications. In this paper, a new algorithm for the detection of melanoma from digital images using multiple extraction capabilities with a support vector machine (SVM) for very precise classification. To train SVM, more characteristics such as color, texture, and statistical characteristics are taken into account. To increase the sensitivity of the classification, feature extraction is performed in the RGB (red, green and blue) and HIS (hue, intensity, saturation) color domains. Two separate modules for training and testing are performed with sample data collected with the help of medical experts. To separate the skin region, a song-based segmentation technique is used in the gray scale component of the input image. The proposed method is tested with various images which are collected from different patients from different locations. From the validation of the result, it is clear that the proposed algorithm can provide a maximum precision of 95%, which is the best compared to conventional classification algorithms

Keywords: Melanoma, RGB, HIS, Support Vector Machine, Structural characteristic, Statistical characteristic.

PDF | DOI: 10.17148/IJARCCE.2021.101006

Open chat
Chat with IJARCCE