Active contour without edges for image segmentation based on selective attention
Hui Li, Yanjiang Wang · 2014
To remedy the shortcoming of the initial contour us ually be given manually in majority of current acti ve contour algorithms, this paper puts forward an improved act ive contour without edges algorithm for image segme ntation based on selective attention. Firstly, the mathemat ical morphology operator is used for smoothing prep rocessing, then the selective attention mechanism is incorporated i nto the active contour model without edges and the wavelet transform is applied to mask initialization. Finall y, the active contour model without edges is applie d for image segmentation. Experimental results show that the al gorithm proposed can not only reduce the number of iterations, but also get more accurate segmentation results for multiple-target image and improve the efficiency o f active contour algorithm obviously.