Abstract :
The visual attention model and the Mean Value Seamless Cloning (MVSC) algorithm are combined to propose an image mosaicing algorithm based on salient region and MVSC. The new algorithm uses the visual attention model to extract the salient region, and to use the regional registration technology to achieve the image matching, and MVSC algorithm possess good image fusion character, so the image mosaics were studied. The traditional visual attention is improved, and using the improved visual attention model algorithm to obtain salient region of image repeated. The salient regions of image repeated are matched by the image registration algorithm to determine clone domain of the source and target. Finally, the clone domains of the source and target are fused by the MVSC algorithm, to achieve image mosaicing. Experimental results show that this paper algorithm not only can automatically and accurately obtain salient regions, but also reduce the complexity of the image registration, improve the quality of image mosaicing.
