A method for image salient regions detection based on layers and dynamic threshold
基于层次和动态阈值的图像显著区域检测方法

The model for detecting salient regions based on layers was developed and consummated. Primary images were divided into layers. And then for every layer, basic characteristics were extracted, and saliency characteristics including intensity saliency, saturation saliency and color saliency were calculated. Because of the possibility characteristics of images, the planar normal distribution function was adopted to present the pertinence of pixels accurately and reasonably. The results of layers were added and the dynamic threshold was used to reach the final saliency regions. Dynamic threshold method can be influenced hardly by the primary images, so it can adapt to different situations. When the method was applied to real images, satisfying results were got.