Salient Object Detection And Its Application Based On Color Feature And Objectness

With the development of Internet,large amount of data increase every day,including the digital image resources.Facing with mass digital data,people need the digital data to be processed by computer intelligently.More and more data need to be compressed,coding and so on to solve the data redundancy.Saliency has been the core part and solution on many image processing task based on image content,such as image retrieval and image compress.On the same time,people are more likely to understand the importance of the image,which needs the help of the computer to solve the image recognition and image understanding by simulate human visual system.Salient object detection has been the important part among the advanced image process task,and become the hot topic today.Traditional saliency measure can only detect the edge and interesting point of the target object without the whole body of the salient object.To solve the problem of traditional saliency measure,we propose a novel saliency method based on color feature.Firstly,pixel-level saliency is obtained by computing the pixel's sparseness in RGB space and the color contrast.Because the CIELab space meets the principle of human visual system,we calculate the color contrast in CIELab space.The image is segmented into super-pixels and the average saliency of all the pixels within a super-pixel is considered as its saliency value.We calculate super-pixel saliency by averaging pixel-level saliency value which is in each super-pixel.An energy function is used to smooth the saliency value of adjacent super-pixels,which is helpful to obtain a smooth saliency map with distinct foreground object.Secondly,to meet the real-time need of the object detection application,we use object proposal to search the target object in the digital image instead of traditional sliding window measure.The classic objectness method is optimized by omitting the independent SR saliency measure and changing the window scoring function by our Bayesian posterior probability.Finally,we propose a novel salient object detection algorithm by fusing color saliency and the optimized objectness.We search the target object by objectness among the color saliency map through the center-surround principal.The fusion result is considered as the prior probability of object-level saliency.The likelihood probability of object-level saliency is obtained by calculating the ratio between the foreground cluster in the objectness window and the total pixels in the window.Finally,an object-level saliency posterior probability is calculated by the Bayesian model.Saliency experiments on MSRA and ECSSD dataset have shown that our saliency method have a better performance and high detection precision than the other eight state-of-the-art algorithms.We also experiment the performance between our algorithm with other three proposal measure,and the proposal windows generated by our method largely cover the foreground object in the color saliency map.Finally,object-level saliency experiments on the dataset show that our method has a better performance than the other similar methods which detect salient object with objectness.Since the model proposed before have a lot of thread hold parameters,we make the experiment to verification the parameters in each model.Such as analyzing the sensitivity of the parameters in the energy function and giving the method of finding the binary thread hold.We also experiment the time complexity of both color saliency and the final salient object detection method,experiment result show that our method meet the real-time need of the application.Finally,we apply our color saliency on the traffic sign detection application;experiments show that our method has a better performance than the traditional saliency method and traffic sign detection method based on HSV space.