Accurate diagnosis of the skin lesion boundaries is a difficult task even for an expert (experienced specialists). Thus automatic segmentation of dermoscopic images has been proposed in recent years. In this paper, a hybrid dermoscopic image segmentation method has been proposed which employs fast fuzzy c-mean, YUV and wavelet decomposition properties. In the proposed method, RGB image is converted to YUV color space and U channel after denoising is enhanced by applying contrast adjustment. Then, a primary segmentation image is obtained using statistical gray level histogram based fast version of fuzzy cmean clustering. The final segmented image would be achieved after employing wavelet transformation, Otsu thresholding and morphological reconstruction algorithms. The experimental results reveal that the proposed scheme not only attains higher accuracy and sensitivity, also has lower false negative and positive errors in compared with the related existing methods.

Full text: IJEIR-1227_final.pdf