Dependency perception object position reconfiguration based image enhancing method

The invention discloses a dependency perception object position reconfiguration based image enhancing method. The method comprises extracting foreground objects of images preliminarily; dividing the images into a plurality of areas, performing foreground object dependency analysis on each area to obtain dependency perception objects; setting dependency perception object position optimization targets according to photographical artistic criteria and position relations among the dependency perception objects, and performing composition optimization; performing background completion on missing parts of optimized images, and performing trimming processing to obtain final image enhancement results. According to the method, dependency perception objects in images can be analyzed and extracted automatically, and analysis results can be used for image configuration optimization.