A novel multiscale model named generalized linear scale autoregressive(GLSA) is proposed,while both approximation and detail information of images are considered.Firstly,GLSA model is established to represent the map between the pixels residing at images of various resolutions;after that,parameters of the model are obtained from original image and its wavelet decomposition results to decide the map;finally,the identified map is used to estimate images at finer resolution from coarser versions.The model is applied to face recognition.Parameters of the model about testing image are compared with that about training images are compared to classify the testing image.Experimental results show that images estimated by GLSA model are more similar to target images than those estimated by LSA model.The GLSA-based face recognition system is robust to illumination.