TY - GEN
T1 - An Improved Image Super-Resolution Algorithm for Percutaneous Endoscopic Lumbar Discectomy
AU - Li, Xue
AU - Zhou, Zihan
AU - Wang, Kaifeng
AU - Liu, Haiying
AU - Qian, Yalong
AU - Duan, Xingguang
AU - Li, Changsheng
N1 - Publisher Copyright:
© 2024, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
PY - 2024
Y1 - 2024
N2 - High-resolution (HR) spinal endoscopic images are essential to enhance the surgeon’s visual presence for the guidance of surgical procedures. However, available image super-resolution methods, especially deep learning methods, are mostly trained with open-source life scene datasets which possess limited medical image features. To address this issue, we have proposed an improved SRGAN model for the visual enhancement of percutaneous endoscopic lumbar discectomy (PELD) surgical images. Specifically, a residual dense block (RDB) and a dynamic RELU function are introduced. We validate the proposed method on PELD datasets. Quantitative and qualitative comparisons are carried out by comparing methods. The method proposed in this paper improves PSNR by 2.8% and SSIM by 6% compared with the original SRGAN, which proves the superiority of this methods.
AB - High-resolution (HR) spinal endoscopic images are essential to enhance the surgeon’s visual presence for the guidance of surgical procedures. However, available image super-resolution methods, especially deep learning methods, are mostly trained with open-source life scene datasets which possess limited medical image features. To address this issue, we have proposed an improved SRGAN model for the visual enhancement of percutaneous endoscopic lumbar discectomy (PELD) surgical images. Specifically, a residual dense block (RDB) and a dynamic RELU function are introduced. We validate the proposed method on PELD datasets. Quantitative and qualitative comparisons are carried out by comparing methods. The method proposed in this paper improves PSNR by 2.8% and SSIM by 6% compared with the original SRGAN, which proves the superiority of this methods.
KW - Generative Adversarial Network
KW - Medical images processing
KW - Percutaneous Endoscopic Lumbar Discectomy
KW - Single Image Super-Resolution
UR - http://www.scopus.com/inward/record.url?scp=85177177099&partnerID=8YFLogxK
U2 - 10.1007/978-981-99-8018-5_11
DO - 10.1007/978-981-99-8018-5_11
M3 - Conference contribution
AN - SCOPUS:85177177099
SN - 9789819980178
T3 - Communications in Computer and Information Science
SP - 149
EP - 160
BT - Cognitive Systems and Information Processing - 8th International Conference, ICCSIP 2023, Revised Selected Papers
A2 - Sun, Fuchun
A2 - Fang, Bin
A2 - Meng, Qinghu
A2 - Fu, Zhumu
PB - Springer Science and Business Media Deutschland GmbH
T2 - 8th International Conference on Cognitive Systems and Information Processing, ICCSIP 2023
Y2 - 10 August 2023 through 12 August 2023
ER -