내시경 수술을 위한 딥러닝 시스템 : 통합형 영상 강화 및 연기 제거
Deep Learning System for Endoscopic Surgery: Integrated Image Enhancement and Smoke Removal
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This study proposes a deep learning-based system to overcome the challenges of degraded image quality due to limited lighting and small size of camera in endoscopic surgery, particularly addressing the issue of smoke obscuring the surgeon’s view. The image enhancement algorithm consists of dark area improvement, blood vessel emphasis, and noise reduction, implemented through filtering and histogram equalization in the Lab color space. For smoke removal, a Generative Adversarial Network-based model was employed, using synthetic data generated by adding artificial smoke via Perlin noise to the original images. The results demonstrate the effectiveness of the pro- posed system, with significant improvements in visual clarity, particularly in dark areas and blood vessel visibility, as evidenced in qualitative and quantitative analyses. Test results demonstrated that the proposed system achieved superior image quality compared to existing smoke-specific removal models, with Peak Signal-to-Noise Ratio and Structural Similarity Index Measure indices of 23.17 and 0.8802, respectively. These outcomes are expected to sig- nificantly contribute to improving the efficacy and safety of endoscopic surgeries.
Abstract
This study proposes a deep learning-based system to overcome the challenges of degraded image quality due to limited lighting and small size of camera in endoscopic surgery, particularly addressing the issue of smoke obscuring the surgeon’s view. The image enhancement algorithm consists of dark area improvement, blood vessel emphasis, and noise reduction, implemented through filtering and histogram equalization in the Lab color space. For smoke removal, a Generative Adversarial Network-based model was employed, using synthetic data generated by adding artificial smoke via Perlin noise to the original images. The results demonstrate the effectiveness of the pro- posed system, with significant improvements in visual clarity, particularly in dark areas and blood vessel visibility, as evidenced in qualitative and quantitative analyses. Test results demonstrated that the proposed system achieved superior image quality compared to existing smoke-specific removal models, with Peak Signal-to-Noise Ratio and Structural Similarity Index Measure indices of 23.17 and 0.8802, respectively. These outcomes are expected to sig- nificantly contribute to improving the efficacy and safety of endoscopic surgeries.
참고문헌 (23)
- Darzi A, "The impact of minimally invasive surgical techniques", Annu Rev Med, 55(1), pp.223~237 (2004)
- Ko WJ, "Image quality analysis of various gastrointestinal endoscopes : why image quality is a prerequisite for proper diagnostic and therapeutic endoscopy", Clin Endosc, 48(5), pp.374~379 (2015)
- Matsuda T, "Advances in image enhancement in colonoscopy for detection of adenomas", Nat Rev Gastro Hepat, 14(5), pp.305~314 (2017)
- Xia W, "Endoscopic image enhancement with noise suppression", Healthc Technol Lett, 5(5), pp.154~157 (2018)
- Ulmer BC, "The hazards of surgical smoke", AORN journal, 87(4), pp.721~738 (2008)
- Togashi K, "A comparison of conventional endoscopy, chromoendoscopy, and the optimal-band imaging system for the differentiation of neoplastic and nonneoplastic colonic polyps", Gastrointest Endosc, 69(3), pp.734~741 (2009)
- Zhang G, "A medical endoscope image enhancement method based on improved weighted guided filtering", Mathematics, 10(9), pp.1423~ (2022)
- Zhang K, "An image enhancement technique using nonlinear transfer function and unsharp masking in multispectral endoscope", International Conference on Innovative Optical Health Science, pp.16~23 (2017)
- Zhang C, "Nonlinear distortion correction in endoscopic video images", Proceedings 2000 International Conference on Image Processing, pp.439~442 (2000)
- Reza AM, "Realization of the contrast limited adaptive histogram equalization (CLAHE) for real-time image enhancement", J VLSI Signal Process Syst Signal Image Video Technol, 38, pp.35~44 (2004)
- He K, "Single image haze removal using dark channel prior", IEEE Trans Pattern Anal Mach Intell, 33(12), pp.2341~2353 (2010)
- Chen W-T, "Color channel-based smoke removal algorithm using machine learning for static images", 2018 25th IEEE International Conference on Image Processing (ICIP), pp.2855~2859 (2018)
- Leibetseder A, "Real-time image-based smoke detection in endoscopic videos", Proceedings of the on Thematic Workshops of ACM Multimedia, pp.296~304 (2017)
- Asif M, "An automatic framework for endoscopic image restoration and enhancement", Appl Intell, 51, pp.1959~1971 (2021)
- Venkatesh V, "Unsupervised smoke to desmoked laparoscopic surgery images using contrast driven Cyclic-DesmokeGAN", Comput Biol Med, 123, pp.103873~ (2020)
- Jiang Y, "Enlightengan : Deep light enhancement without paired supervision", IEEE Trans Image Process, 30, pp.2340~2349 (2021)
- Salazar-Colores S, "Desmoking laparoscopy surgery images using an imageto-image translation guided by an embedded dark channel", IEEE Access, 8, pp.208898~208909 (2020)
- Pan Y, "DeSmoke-LAP : improved unpaired image-to-image translation for desmoking in laparoscopic surgery", Int J Comput Assist Radiol Surg, 17(5), pp.885~893 (2022)
- Perona P, "Scale-space and edge detection using anisotropic diffusion", IEEE Trans Pattern Anal Mach Intell, 12(7), pp.629~639 (1990)
- Goodfellow I, "Generative adversarial networks", Commun ACM, 63(11), pp.139~144 (2020)
- Isola P, "Image-to-image translation with conditional adversarial networks", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.1125~1134 (2017)
- Ronneberger O, "U-net : Convolutional networks for biomedical image segmentation", MICCAI 2015:18th international conference, pp.234~241 (2015)
- Perlin K, "An image synthesizer", ACM Siggraph Computer Graphics, 19(3), pp.287~296 (1985)