Abstract: An algorithm for multi-meanings image synthesis which based on wavelet transform is presented. Firstly，the two target images were decomposed by wavelet transform with the scale which was selected by visual angles. Secondly, the multi-meanings image viewed in two angles was reconstructed by elective fusion algorithms and variable weight. The algorithm based on wavelet transform and two existing algorithms were compared and analyzed. Experimental results show that the images produced by this algorithm had better characteristics of multi-meanings and compelling visible effect，and contained sufficient color and details of original images. Finally, the results were averaged for denoising.

Key words: computer image processing, wavelet transform, feature of human vision, image fusion, multi-meanings image, target image