Dynamic Range Compression Algorithm Based on Green’s Function

Due to the uncomfortable visual effect and unclearly details which exist in high dynamic range images, this paper proposes a dynamic range compression algorithm based on Green's function. This algorithm employs the sparse representation method to obtain irregular regions which possess wide disparity in term of luminance.The similar regions of shadow and highlight domains are retrieved by calculating of the earth mover distance between regions.A numerical solution for irregular regions is introduced to calculate low dynamic range images which are derived from the curvilinear integral according to the Dirichlet boundary condition. Experimental results show that the sparse representation method restrains the accumulative errors which are caused by curvilinear integral.The numerical solution enhances the image details efficiently, meanwhile the dynamic range is compressed. The resulting images represent excellent effect of sense and high clarity.