| 摘要 |
Disclosed are an image segmentation method and device, which belong to the field of image processing. The image segmentation method comprises: establishing a saliency model of an image; acquiring a foreground sample point and a background sample point in the image according to the saliency model; establishing a foreground and a background classification model according to the saliency model and the foreground sample point and the background sample point; and segmenting the image according to a predetermined image segmentation algorithm, the predetermined image segmentation algorithm segmenting the image using the foreground and the background classification model and edge information between pixel points. By automatically determining the foreground and the background sample point and establishing the foreground and the background classification model with the aid of the saliency model, image segmentation is realized using the foreground and the background classification model; and the problem in the related art that the segmentation efficiency is relatively low when a lot of images are segmented due to the fact that a user must roughly select a foreground sample point and a background sample point manually is solved; thereby achieving the effects of being able to automatically select samples, thus increasing the classification accuracy. |
