Research On Visual Saliency Target Detection Technology Based On High-Level Semantics

Neurobiologists and cognitive psychologists have shown that human vision has a quick and easy to grasp the interesting and conspicuous goals in visual observation scenes.In order to enable computers to acquire this capability,more and more researchers have studied visual saliency target detection technology.Research on visual saliency target detection technology has become a hot research topic in the field of computer vision and image recognition.The purpose of visual saliency target detection technology is to determine the saliency regions of the image that can attract human visual attention,rather than just detecting several salient features in the image,so it can not only accelerate the visual process,but also improve the visual task performance.Therefore,the research of visual saliency target detection technology has far reaching research significance.Although the current research on visual saliency detection technology has already had certain research results,when the image background is complex or the target is too much,the problem of the accuracy and efficiency of the saliency target detection result still exists.In view of the above description,this paper will introduce and research the visual saliency target detection technology.The main work and research innovations of this paper are as follows:(1)Visual saliency target detection technology has been widely applied in recent years,making it an important contribution in the field of image recognition.This paper will discuss the relevant theoretical knowledge of visual saliency target detection technology in detail,and introduce how the visual saliency target detection technology can efficiently acquire image target information.At the same time,analyze and summarize some classic popular saliency detection algorithms.And through the experimental simulation to analyze the performance of the algorithm,and finally summarize the key issues that should be paid attention to when constructing the saliency detection algorithm model.(2)Aiming at the problem that the background color of the image and the visual saliency detection target image are similar,which leads to the accuracy and efficiency of the saliency detection result,a diffusion-based visual saliency target detection model is proposed.In this paper,the diffusion matrix is recombined by constructing the symmetric matrix and discarding the constant eigenvectors and discriminating power.At the same time,the seed vector in the diffusion matrix is reconstructed.Finally,the recombined diffusion matrix and the reconstructed seed vector are diffused.A diffusion-based visual saliency target detection algorithm.Through the experimental simulation comparison,the algorithm has some improvement compared with the previous algorithm.When the background color and the visual target information are similar in color,the saliency target detection effect of the algorithm is obviously improved.(3)Aiming at the problem that the low-level features of the image can not uniformly detect the saliency target,this paper combines the high-level a priori semantics with the low-level image features,and proposes a novel algorithm based on high-level prior semantics.By using the deep convolutional neural network to segment and extract the semantics of the image,the explicit prior information is semantically segmented and extracted to obtain the explicit saliency detection map;then the implicit prior saliency features and saliency values in the image are performed.The trained training model is used to calculate the implicit saliency map.Finally,the explicit saliency detection map and the implicit saliency detection map are adaptively fused to form a saliency detection map that uniformly covers the saliency target pixels.A large number of experimental results show that compared with other algorithms,the saliency target detection effect of the algorithm is significantly improved compared with other methods.Finally,the challenges still need to be faced in the emergence of saliency target detection algorithms,and the future research trends of saliency target detection algorithms are discussed.