The easy path wavelet transform (EPWT) has recently been proposed for representations of discrete data, in particular for image data. This paper presents a new hybrid method for image approximation that exploits the advantages of the usual tensor product wavelet transform for the representation of smooth images and uses the EPWT for an efficient representation of edges and texture. Predetermined amount of pixel coefficients are taken by hard thresholding procedure. This method provides the better approximation results. However the EPWT suffers from its adaptivity costs that arise from the storage of path vector. It exploits the local correlation of the neighboring pixel and best neighbor strategy is adapted.