Interactive Image Editing By Dense Labeling

Interactive image editing generally uses an image editing mode that delineates specific image regions and assigns specific physical meanings to these selection operations.The current interactive editing technology relies on a large number of precise user choices,it is difficult to achieve real-time feedback in terms of operational efficiency,requiring additional resources when solving complex image editing problems,which raise the application threshold of complex image processing technology.In order to reduce the threshold of complex image editing,simplify the implementation process of interactive image editing,and improve the feedback efficiency of editing,this paper proposes an image editing mode based on dense labeling,and applies this model to the effect of editing depth of field and image defogging.In complex image processing applications.This algorithm combines the efficiency of super-pixel processing and the accuracy of per-pixel processing.It employs a linear system to express and solve image feature information and userassisted operations,which reduces the requirements for the user's professional experience and improves the speed of response and feedback.In this paper we present three complex image processing applications: DoF,dehazing and tone-mapping.The specific physical model and the combination of effective user operation methods are applied to the proposed algorithm.Complex image processing operations can be performed without additional physical information...