Article
Super-Resolution Network with Information Distillation and
Multi-Scale Attention for Medical CT Image
Tianliu Zhao
1
, Lei Hu
1,
* , Yongmei Zhang
2
and Jianying Fang
1
Citation: Zhao, T.; Hu, L.; Zhang, Y.;
Fang, J. Super-Resolution Network
with Information Distillation and
Multi-Scale Attention for Medical CT
Image. Sensors 2021, 21, 6870.
https://doi.org/10.3390/s21206870
Academic Editor: Nunzio Cennamo
Received: 1 September 2021
Accepted: 12 October 2021
Published: 16 October 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
1
School of Computer Information Engineering, Jiangxi Normal University, Nanchang 330022, China;
tianliuzhao@jxnu.edu.cn (T.Z.); jyfang@jxnu.edu.cn (J.F.)
2
School of Information Science and Technology, North China University of Technology, Beijing 100144, China;
zhangym@ncut.edu.cn
* Correspondence: hulei@jxnu.edu.cn
Abstract:
The CT image is an important reference for clinical diagnosis. However, due to the external
influence and equipment limitation in the imaging, the CT image often has problems such as blurring,
a lack of detail and unclear edges, which affect the subsequent diagnosis. In order to obtain high-
quality medical CT images, we propose an information distillation and multi-scale attention network
(IDMAN) for medical CT image super-resolution reconstruction. In a deep residual network, instead
of only adding the convolution layer repeatedly, we introduce information distillation to make full
use of the feature information. In addition, in order to better capture information and focus on
more important features, we use a multi-scale attention block with multiple branches, which can
automatically generate weights to adjust the network. Through these improvements, our model
effectively solves the problems of insufficient feature utilization and single attention source, improves
the learning ability and expression ability, and thus can reconstruct the higher quality medical
CT image. We conduct a series of experiments; the results show that our method outperforms the
previous algorithms and has a better performance of medical CT image reconstruction in the objective
evaluation and visual effect.
Keywords:
super-resolution; medical CT image; multi-scale attention; information distillation;
deep learning
1. Introduction
The computed tomography (CT) image is an important auxiliary means in clinical
diagnosis. The image quality has a very significant impact on the diagnosis of lesions. High-
quality medical images can help doctors to identify the symptoms more accurately and
quickly formulate the corresponding treatment plan for patients. However, the limitation
of imaging devices makes it difficult to obtain high-resolution medical CT images, so these
images always have some problems such as low resolution, blurring and loss of detail. As
a classic computer vision task, super-resolution (SR) reconstruction can use low-resolution
(LR) images to reconstruct high-resolution (HR) images. Super-resolution algorithms can
also be used in medical CT image to improve the image quality.
According to the different objects of SR processing, we can divide the super-resolution
technique into single image super-resolution (SISR), multiple image super-resolution
(MISR) and video super-resolution (VSR). Among them, SISR only uses one image to
improve the resolution of the image. The requirement of input is relatively low, so there
are many studies regarding SISR. SISR is one of the key research directions on image
super-resolution. With the improvement performance of the super-resolution algorithm,
the super-resolution network is gradually applied in the field of medical image. In this pa-
per, we mainly discussed the single image super-resolution reconstruction for the medical
CT image.
Sensors 2021, 21, 6870. https://doi.org/10.3390/s21206870 https://www.mdpi.com/journal/sensors
