Abstract: The most challenging issue for content-aware image retargeting is how to preserve the important content and minimize the visual distortions on the image structures.In this paper,we present an improved method for content-aware image resizing by leveraging a global optimization process.We reformulate the problem of image retargeting to a mesh deformation task and build a global shape preserving energy function on the grid.By solving a convex quadratic program with constraints,the deformed mesh coordinates are obtained,yielding the final retargeting image.Experimental results demonstrate that this approach not only preserves the important content and feature structures,but also preserves the global image configuration with low visual distortion.