abstract |
The present invention provides a saliency map fusion method for community images, which includes inputting a training image, including the following steps, for image I in D, applying m kinds of extraction methods to extract the saliency map of the training image, wherein D is training Set; Calculate AUC value; According to the calculation method of step 1 and step 2, obtain the sorting list of the extraction method of every image, the sorting list collection is T ; The neighbor search that carries out in training set; The result in step 4 is merged; Saliency maps for fused test images. The purpose of the saliency map fusion method for community images proposed by the present invention is to propose a specific saliency map fusion method for the characteristics of community images, and the performance of fusion is greatly improved compared with that of a single method before fusion. |