Improved Model of Salient Regions Extraction
PI Yan-qin · Video Engineering · 2012
A novel algorithm for salient region extraction is presented.First,four characteristics graphs of edging,brightness,color,the orientation of the original image are extracted respectively.Then,multi-scale image features of the four characters are combined into one saliency map.In orientation feature extraction,reciprocal function is used to make the orientation feature map more complete than using Gabor function.The algorithm proposed in this paper is faster and easier to realize and can extract more obvious salient maps than Itti model.