Title :
Level set segmentation of dermoscopy images
Author :
Silveira, Margarida ; Marques, Jorge S.
Author_Institution :
Inst. de Sist. e Robot., Inst. Super. Tecnico, Lisbon
Abstract :
This paper presents a method for the segmentation of skin lesions in dermoscopy images. The proposed technique uses region based level sets and adopts a mixture of Gaussian densities as the probabilistic model for both the lesion and the skin regions. The estimation-maximization (EM) algorithm is used to estimate the probability density functions in each region. Experimental results which demonstrate the effectiveness of the proposed technique are presented.
Keywords :
Gaussian processes; density functional theory; expectation-maximisation algorithm; image segmentation; medical image processing; skin; Gaussian density; dermoscopy image; estimation maximization algorithm; level set segmentation; probabilistic model; probability density function; skin lesion; Active contours; Clustering algorithms; Image segmentation; In vivo; Lesions; Level set; Pigmentation; Probability density function; Region 1; Skin; EM-algorithm; dermoscopy; level sets; segmentation;
Conference_Titel :
Biomedical Imaging: From Nano to Macro, 2008. ISBI 2008. 5th IEEE International Symposium on
Conference_Location :
Paris
Print_ISBN :
978-1-4244-2002-5
Electronic_ISBN :
978-1-4244-2003-2
DOI :
10.1109/ISBI.2008.4540960
