Extracting Salient Features Of Images Based On Region Contrast Optimization

Now the feature extraction of images is widely used in different fields,and playing a very important role.Significant processing plays an important role in computer vision and can be applied in different areas in this field.At the same time,the combination of significant feature extraction can achieve better extraction effect,in practice we can combined with the actual situation,and explore more valuable significant treatment.Detection and segmentation of salient objects in natural scenes of everyday life,known as significant target detection,has attracted many focused studies in computer vision,and has resulted in a number of practical applications.However,despite the existence of many models in life,but for the existing results and problems encountered in the in-depth understanding and use is still very lacking.In addition,significant features are applied to different disciplines,including biological sciences and pharmacy,and significant feature extraction based on significance detection can promote the development of subject technology.Therefore,it is of great value to study the significant feature extraction of images.At present,there is a more mature significant detection algorithm,but with the application of the scope of the increase,the new algorithm used to solve the problems encountered should also be gradually optimized.In this paper,the original image is mapped to the feature space,and the original image is divided.And then according to the partition area of the European distance and weight to calculate the significance value,thus determining the significant area of the shape of the boundary and other factors.In this paper,the optimization of the realization of the part is mainly for Gaussian difference weighted optimization.Besides,we have studied the existing algorithm of image salience,and analyzed these classic algorithms about the analysis of the existing regional comparison.Through theanalysis of global contrast and region contrast,results based on the contrast of image are put forward,and the improved aspect combines the region contrast characteristic and global contrast characteristic of the image.The GC and RC algorithms are used in the process of extracting the salient features of images,and the weighting of the spatial information is used in the process of the completion.In this paper,we complete the calculation based on the Gaussian Distribution Model,and obtain the center position of the center,that is the vertical and horizontal coordinates,we can calculate the distance between the pixel point and the central area.And the region contrast algorithm is used to optimize the spatial weighting of the region contrast algorithm.The difference between the significant target and the surrounding area is increased,the difference denoising of the Gaussian function is used,and the difference of each Gaussian function is weighted,and different weights are set to realize enhance the information of low frequency region,and the weight set for the lower part of the frequency is greater.Finally,through the experimental verification,we can get the experimental results,we can extract to a better target.It is seen from the salient diagram of the experimental results that the target complete extraction of the high contrast region is achieved,the significant value of the surrounding area of the target is reduced,and the region with significant target can be better exhibited.