Research On Image Significance Detection Based On Regional Comparison

There is an old saying: seeing is believing.This statement fully reflects the importance of human vision in the understanding and perception of the world.The eyes,as the first window to the human experience of the surrounding environment,are mostly obtained from the eyes.With the development of computer technology,human beings try to endow the computer with the ability of human visual system.In the fields of computer vision and pattern recognition,human visual information is embodied in the analysis and understanding of images.The human eye receives a vast amount of visual information every day,but not all of it is processed by the visual system and the brain.Only information that is interesting and important to humans will be left behind,and the rest will be ignored.In the field of computer vision and pattern recognition and so on,hope can through the significance of the image detection,extract significant objects in the image,highlight the significant section of the image and get the significant figure images.The significance of image detection is in the fields of biology,psychology and computer science.In the course of more than 20 years of research,there have been a lot of image significance detection models proposed.However,with the deepening of the research,researchers also face some difficult problems that need to be solved urgently.The significant detection of images is applied to many fields,such as medical research,biology and industry,and presents a broad commercial prospect.Therefore,it is of great significance to study it.In this article,based on the current image significance detection theory of knowledge learning and research,from the human visual characteristics,visual significance,theoretical basis and significant features of related image significance detection are summarized in terms of the basic knowledge,and the significance of research and analysis of several typical detection algorithm.As image segmentation is an important part of image significance detection,it plays an important role in the final image significance.In SLIC super pixel segmentation algorithm on the basis of image segmentation on multi-scale,the edge information of image is good,in the subsequent process had good effect on significant value calculation,in a significant figure can be more clearly reflects the significant objects.The global contrast of image segmentation at each scale was calculated and the average value of multi-scale significance value was taken as the significance graph.The significant graph obtained by the HC algorithm based on global contrast is combined with the significant graph obtained by the AC algorithm based on the local contrast,and the significant graph of the combination is obtained by comparing the two regions.In addition,significant regional focus processing was carried out for the combined significant figure,and the significance of the significant region was enhanced,increasing the difference with the surrounding area and highlighting the significant objects from the surrounding environment.Finally,the practical application of the algorithm is obtained,and the optimized experimental results are obtained.From the comparison between the experimental results and the significant graphs obtained from other algorithms,it can be seen that the significant regions show a better significance and can better highlight the significant objects.