As a low-frequency and smooth signal, the bias field has a certain destructive effect on magnetic resonance(MR) images, and is the main obstacle for doctors' diagnosis and image processing (such as segmentation, texture analysis, and registration). Before analyzing a damaged MR image, a preprocessing step is needed to correct the bias field in the image. Unlike traditional bias field removal algorithms based on signal models and a priori assumptions, deep learning methods do not require precisely modeling signals and bias fields, and do not need to adjust parameters. After the large training set is trained through the deep neural network, the MR image with the bias field is input, and the corrected MR image is output. In this paper, we propose taking the local feature images of the bias field in multiple frequency bands obtained by the log-Gabor filter bank and the original image as input, correct the bias field of a brain magnetic resonance image through a deep separable convolutional neural network, and use residual learning and batch normalization to speed up the training process and improve bias field correction performance. Our training model was tested on the BrainWeb simulated database and HCP real dataset, and the results of the qualitative analysis show that our training model achieved better performance than the traditional state-of-the-art N4 and NIMS (Non-Iterative Multi-Scale) methods.