This paper presents a novel way to reduce noise introduced by different image enhancement techniques. As the human visual system is highly sensitive to change in brightness, the proposed method is applied to the luma channel of both the non-enhanced and enhanced image. The basic assumption is that the non-enhanced image is either free of noise or noise is present but not perceivable. In order to avoid inappropriate assumptions on the statistical characteristics of noise, a different one is proposed. Also, it gives the importance of directional content in human vision and the analysis is performed through the Dual-Tree Complex Wavelet Transform (DTCWT). Compared to discrete wavelet transform, the DTCWT provides distinction of data directionality in the transform space. The standard deviation for each level of the transform of an non-enhanced image coefficients is computed and normalized across the six orientations of the DTWCT. The normalized said map is then used to shrink the coefficients of the enhanced image. The shrunk coefficients of the enhanced image are mixed according to data directionality, with the coefficients of non-enhanced image. Finally, the inverse transform provides the noise-reduced version. The proposed one thoroughly reduces the noise introduced by the enhancement methods and produces better improvement in PSNR of the image. In order to confirm the validity of the proposed method, a through numerical analysis of the results has been done.