DCT-FANet: DCT based frequency attention network for single image super-resolution

In single-image super-resolution (SISR) task, it is challenging to recover high-frequency details from a low-resolution (LR) image due to its ill-posed problem. Most existing CNN-based super-resolution (SR) methods pursue an end-to-end solution from the low-resolution image to the high-resolution image through deeper or wider networks. However, these methods treat high-frequency information and low-frequency information equally which hinder the recovery of high-frequency details and result in low visual comfort. In this paper, we propose a DCT based frequency attention network (DCT-FANet) to distinguish different frequency of LR images, and enhance the high-frequency information adaptively. Specially, a DCT spatial cube extraction (DCT-SCE) module is proposed to decompose LR images into multiple frequency subbands. Different from other DCT based SISR methods that use DCT coefficients as subbands, our module considers the feature difference between spatial domain and frequency domain, and remains spatial structure information of each DCT subband. Then, based on the characteristics of DCT subbands, we design a Gaussian based frequency selection (GFS) module to give more attention to the high-frequency information. To promote communication between each frequency subband, an adaptive non-local dual attention (ANDA) module is developed, leading to a further enhancement of high-frequency information and improvement of the final performance. Experimental results demonstrates that our DCT-FANet recovers more high-frequency details, and achieves excellent performance with fewer parameters. •We construct a DCT-FANet to reconstruct more image details for SISR. Experimental results demonstrate that our model can recover more high-frequency details, and achieve excellent performance with fewer parameters.•A DCT-SCE module is developed to decompose the LR image into multiple frequency subbands. This module bridges the features between spatial domain and frequency domain, which can remain spatial structure information of each DCT subband, hence improving the scalability of extracted DCT subbands. To the best of our knowledge, this is the first time to consider the feature difference between spatial domain and frequency domain in deep-learning-based SISR task.•To make full use of limited high-frequency information, we explore the characteristics of extracted DCT subbands, and propose a GFS module. This module can give high-frequency subbands more attention according to th