Journal of the Korea Institute of Information and Communication Engineering (한국정보통신학회논문지)

The Korea Institute of Information and Commucation Engineering (한국정보통신학회)
권호 표지
DOI QR코드

DOI QR Code

Structure, Method, and Improved Performance Evaluation Function of SRCNN and VDSR

SRCNN과 VDSR의 구조와 방법 및 개선된 성능평가 함수

  • Lee, Kwang-Chan (School of Computer Information & Communication Engineering, Kunsan National University) ;
  • Wang, Guangxing (Information Technology Center, Jiujiang University) ;
  • Shin, Seong-Yoon (School of Computer Information & Communication Engineering, Kunsan National University)
  • Received : 2021.02.17
  • Accepted : 2021.03.03
  • Published : 2021.04.30
Download PDF
⟨ Previous Next ⟩

Abstract

The higher the resolution of the image, the higher the satisfaction of the viewers of the image, and the super-resolution imaging has a considerable increase in research value among the fields of computer vision and image processing. In this study, the main features of low-resolution image LR are extracted mainly using deep learning super-resolution models. It learns and reconstructs the extracted features, and focuses on reconstruction-based algorithms that generate high-resolution image HR. In this paper, we investigate SRCNN and VDSR in a super-resolution algorithm model based on reconstruction. The structure and algorithm process of the SRCNN and VDSR model are briefly introduced, and the multi-channel and special form are also examined in the improved performance evaluation function, and understand the performance of each algorithm through experiments. In the experiment, an experiment was performed to compare the results of the SRCNN and VDSR models with the peak signal-to-noise ratio and image structure similarity, so that the results can be easily judged.

이미지는 해상도가 높을수록 이미지를 시청하는 사람들의 만족도가 높아지며 초고해상도 이미지화는 컴퓨터 비전이나 영상처리 분야 중에서도 연구 가치가 꽤 높아지고 있다. 본 연구에서는 주로 딥 러닝 초 해상도 모델을 사용하여 저해상도 이미지 LR의 주요 특징을 추출한다. 추출된 특징을 학습 및 재구성하고, 고해상도 이미지 HR을 생성하는 재구성 기반 알고리즘에 중점을 둔다. 본 논문에서는 재구성에 기반을 둔 초 해상도 알고리즘 모델에서 SRCNN과 VDSR에 대하여 알아보도록 한다. SRCNN과 VDSR모델의 구조 및 알고리즘 프로세스를 간략하게 소개하고 개선된 성능평가 함수에서도 다중 채널과 특수한 형태에 대하여 알아보도록 하며, 실험을 통하여 각 알고리즘의 성능을 이해하도록 한다. 실험에서는 SRCNN 및 VDSR 모델의 결과와 피크 신호 대 잡음 비 및 이미지 구조 유사도를 비교하는 실험을 수행하여 결과를 한눈에 볼 수 있도록 하였다.

Keywords

References

  1. Y. L. Seo and S. J. Kang, "Current status and latest trend of deep learning-based super resolution technology," Broadcasting and Media Magazine, vol. 25, no. 2, pp. 7-16, Apr. 2020.
  2. C. Dong, C. Loy, K. He, and X. Tang, "Learning a deep convolutional network for image super-resolution," In Proceedings of the European conference on computer vision, Glasgow, United Kingdom, Springer, Cham, pp. 184-199, Sep. 2014.
  3. J. Kim, J. K. Lee, and K. M. Lee, "Accurate image super-resolution using very deep convolutional networks," In Proceedings of the IEEE conference on computer vision and pattern recognition, Las Vegas, NV, USA, pp. 1646-1654, 2016.
  4. Y. Zhang, Y. Tian, Y. Kong, B. Zhong, and Y. Fu, "Residual dense network for image super-resolution," In Proceedings of the IEEE conference on computer vision and pattern recognition, Salt Lake City, UT, USA, pp. 2472-2481, 2018.
  5. Y. Zhang, K. Li, K. Li, L. Wang, B. Zhong, and Y. Fu, "Image super-resolution using very deep residual channel attention networks," In Proceedings of the European Conference on Computer Vision(ECCV), Munich, Germany, pp. 286-301, 2018.
  6. N. Ahn, B. Kang, and K. A. Sohn, "Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network," In Proceedings of the European Conference on Computer Vision(ECCV), Munich, Germany, pp. 256-272, 2018.
  7. D. Vint, G. Di Caterina, J. J. Soraghan, R. A. Lamb, and D. Humphreys, "Evaluation of performance of VDSR super resolution on real and synthetic images," In 2019 Sensor Signal Processing for Defence Conference(SSPD), Brighton, United Kingdom, pp. 1-5, May. 2019.
  8. Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, "Image quality assessment: From error visibility to structural similarity," IEEE Trans. Image Process., vol. 13, no. 4, pp. 600-612, Apr. 2004. https://doi.org/10.1109/TIP.2003.819861