Considering the fact that human eyes pay more attention to the salient regions of images,this paper proposes a full-reference image quality assessment based on salient region map a nd fidelity map.Firstly,the fidelity map of distorted image is obtained by calculating the similarity betwee n the distorted and undistorted images on brightness and chromaticity components;then,the salient map of the u ndistorted image is obtained by segmenting regional blocks,extracting the global saliency of each block and complementing the edge texture of local feature;finally,using the saliency map as weight,the objective asses sment score is obtained by calculating the weighted fidelity map.The experimental results in LIVE image dat abase show that the image quality assessment method based on saliency detection proposed here has good performanc e on assessing the five kinds of distorted images.