Abstract: As the estimation of background luminance at the mutation of light is usually inaccurate, the traditional Retinex enhancement algorithms suffer from the halo phenomenon, as well as the loss of details and color distortion in the enhanced image. Based on the characteristics of the human visual system, a novel color image enhancement algorithm is proposed in this work. As the human visual system is sensitive to structural features and color information of image, a color-bilateral filter is constructed to estimate the background luminance, which can effectively overcome the halo phenomenon. Moreover, by using the local self-adjustment characteristic of the human visual system, a local contrast enhancement function is introduced to adaptively adjust the intensity of each pixel, to overcome the dilemma that the overall contrast is improved but the local contract is reduced. Finally, a color restoration process is utilized to convert the enhanced intensity image back to the color image. Experimental results show that the proposed algorithm is more effective in terms of visual effects, and the enhanced image is not only more detail-preserving, but it is also more colorful and natural, compared with other methods such as MSRCR.