E-mail Alert
RSS
-
摘要:
针对内窥镜图像中因光照不充分、不均匀而造成的细节模糊问题，提出了一种用于人体上消化道内窥镜图像对比度和亮度增强的算法。通过对自适应伽马校正亮度增强算法和有限对比度自适应直方图均衡化算法改进并进行线性融合。通过对输入图像分别进行亮度增强和对比度增强处理，最终得到线性融合增强图像。将提出的算法应用于开源数据集中的上消化道胃部组织图像，并与现有算法进行了对比，采用峰值信噪比(PSNR)、结构相似度(SSIM)和自然图像质量评价(NIQE)作为图像评价指标。实验结果表明，所提出的图像增强算法与现有算法相比，提高了图像质量，为医疗诊断提供更多的细节信息。
Abstract:An image contrast and brightness enhancement algorithm for human upper gastrointestinal endoscopy is proposed to address the problem of blurring of details such as insufficient and uneven illumination in endoscopic images. The algorithm improves and weighted fusion of the adaptive gamma-corrected luminance enhancement algorithm and contrast-limited adaptive histogram equalization algorithm. The input images are processed separately and the final weighted fused enhanced image is obtained. The proposed algorithm is applied to the partial images of the upper gastrointestinal tract in the open access dataset and compared with the existing algorithms for algorithm effect testing experiments, using peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and natural image quality evaluator (NIQE) as the image evaluation metrics. The experimental results show that the proposed algorithm enhances the image with higher quality than other algorithms, which significantly improves the image quality and provides a good basis for image detection.
-
Overview: In the field of medical imaging, human upper gastrointestinal (GI) endoscopy plays a crucial role in diagnosing and managing various pathologies. However, the diagnostic efficacy of this minimally invasive procedure is often hindered by suboptimal imaging conditions, such as inadequate and irregular illumination, leading to blurred visual details. These challenges underscore the necessity for advanced image enhancement techniques that can effectively address such issues and consequently enhance clinical decision-making. This study aims to propose an innovative algorithm for enhancing image contrast and brightness specifically designed for upper GI endoscopy. Recognizing the shortcomings of current methods in dealing with complex endoscopic images, our research focuses on developing a solution that addresses the dual problems of insufficient and uneven illumination. Our goal is to enhance the visibility of critical anatomical structures without introducing artifacts. Our method innovatively integrates adaptive gamma correction for luminance enhancement with a contrast-limited adaptive histogram equalization (CLAHE) algorithm. Applying these techniques separately to the input images and then performing a weighted fusion, our approach achieves a balanced optimization of image contrast and brightness. This fusion strategy ensures that important image details are preserved while mitigating potential issues such as over-enhancement and noise enhancement that may be associated with individual algorithms. To rigorously evaluate the performance of our proposed algorithm, a series of experiments were conducted on a subset of upper gastrointestinal (GI) images from an open-access dataset. The evaluation included comparisons with several established enhancement algorithms using quantitative metrics such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and natural image quality evaluator (NIQE). The empirical results showed that our algorithm consistently outperformed existing methods on these metrics, demonstrating its superior ability to enhance image quality. Specifically, it achieved higher PSNR values, indicating reduced noise and distortion, and improved SSIM values, reflecting better structural preservation similar to the original image. Furthermore, the decreased NIQE scores validated the naturalness and perceptual quality of the enhanced images. In conclusion, this research introduces a novel and effective image enhancement algorithm for upper GI endoscopy that effectively tackles the common issue of insufficient and inconsistent illumination. The proven ability of this technology to enhance image quality without compromising diagnostic integrity paves the way for more accurate and efficient endoscopic examinations, reinforcing its importance as a cornerstone in the advancement of gastrointestinal diagnostic imaging.
-
-
图 2 (a)原始图像 ；(b)第一层亮度增强图像； (c)第二层CLAHE图像； (d)第三层线性融合图像
Figure 2. (a) Origin; (b) Layer1 luminance enhancement algorithm; (c) Layer 2 CLAHE algorithm; (d) Layer 3 linear fusion algorithm
图 3 Pylorus三组图像(a-c)的算法对比图
Figure 3. Comparison of different algorithms for three sets of images (a-c) of Pylorus
图 4 Retroflex-stomach三组图像(a-c)的算法对比图
Figure 4. Comparison of different algorithms for three sets of images (a-c) of Retroflex-stomach
图 5 Z-line三组图像(a-c)的算法对比图
Figure 5. Comparison of different algorithms for three sets of images (a-c) of Z-line
表 1 数据集HyperKvasir部分内容
Table 1. The contents of the HyperKvasir partial dataset
1st layer 2nd layer 3rd layer No. of images Upper GI Anatomical landmarks Pylorus 999 Retroflex-stomach 764 Z-line 932 Pathological findings Barrett's 41 Barrett's Short Segments 53 Esophagitis Grage-A 403 Esophagitis Grage-B-D 260 
下载: 导出CSV
表 2 NIQE
Table 2. NIQE
图像 原始图像 CLAHE GC AGCWD Retinex-Net Zero-DCE ENDOIMLE 本文算法 Pylorus 3.776 3.371 3.762 3.586 3.626 3.893 3.705 3.543 Retroflex 3.538 3.385 3.802 3.569 3.789 4.265 3.628 3.373 Z-line 3.636 3.069 3.273 3.471 3.821 4.523 3.675 3.112
下载: 导出CSV
表 3 PSNR
Table 3. PSNR
图像 CLAHE GC AGCWD Retinex-Net Zero-DCE ENDOIMLE 本文算法 Pylorus 17.18 8.23 13.45 17.61 16.37 14.34 19.43 Retroflex 12.99 9.15 13.91 18.51 17.61 15.41 18.96 Z-line 13.77 9.77 14.98 18.77 18.03 15.45 19.56
下载: 导出CSV
表 4 SSIM
Table 4. SSIM
图像 CLAHE GC AGCWD Retinex-Net Zero-DCE ENDOIMLE 本文算法 Pylorus 0.938 0.655 0.896 0.711 0.787 0.881 0.955 Retroflex 0.893 0.653 0.894 0.778 0.737 0.884 0.936 Z-line 0.913 0.751 0.901 0.793 0.764 0.871 0.946
下载: 导出CSV
表 5 消融实验
Table 5. Ablation experiment
GC (未线性拉伸) 自适应GC 改进CLAHE 本文算法 PSNR 8.56 10.23 15.68 19.55 SSIM 0.579 0.773 0.897 0.948 NIQE 3.785 3.368 3.186 3.246
下载: 导出CSV
表 6 计算时间
Table 6. Run time
算法 CLAHE GC AGCWD Retinex-Net Zero-DCE ENDOIMLE 本文算法 时间/s 40.35 12.18 6.21 3.45 0.01 5.46 4.56
下载: 导出CSV
-
[1] Aghanouri M, Ghaffari A, Serej N D, et al. New image-guided method for localisation of an active capsule endoscope in the stomach[J]. IET Image Process, 2019, 13(12): 2321−2327. doi: 10.1049/iet-ipr.2018.6366
[2] Omisore O M, Han S P, Xiong J, et al. A review on flexible robotic systems for minimally invasive surgery[J]. IEEE Trans Syst Man Cybern Syst, 2022, 52(1): 631−644. doi: 10.1109/TSMC.2020.3026174
[3] 马龙, 马腾宇, 刘日升. 低光照图像增强算法综述[J]. 中国图象图形学报, 2022, 27(5): 1392−1409. doi: 10.11834/jig.210852
Ma L, Ma T Y, Liu R S. The review of low-light image enhancement[J]. J Image Graphics, 2022, 27(5): 1392−1409. doi: 10.11834/jig.210852
[4] Horiuchi Y, Hirasawa T, Ishizuka N, et al. Diagnostic performance in gastric cancer is higher using endocytoscopy with narrow-band imaging than using magnifying endoscopy with narrow-band imaging[J]. Gastric Cancer, 2021, 24: 417−427. doi: 10.1007/s10120-020-01125-w
[5] Okuhata H, Nakamura H, Hara S, et al. Application of the real-time Retinex image enhancement for endoscopic images[C]//Proceedings of the 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, Osaka, Japan, 2013: 3407–3410. https://doi.org/10.1109/EMBC.2013.6610273.
[6] Xia W Y, Chen E C S, Peters T. Endoscopic image enhancement with noise suppression[J]. Healthcare Technol Lett, 2018, 5(5): 154−157. doi: 10.1049/htl.2018.5067
[7] Cho W H, Hong K S. Extending dynamic range of two color images under different exposures[C]//Proceedings of the 17th International Conference on Pattern Recognition, Cambridge, UK, 2004: 853–856. https://doi.org/10.1109/ICPR.2004.1333906.
[8] Sdiri B, Kaaniche M, Cheikh F A, et al. Efficient enhancement of stereo endoscopic images based on joint wavelet decomposition and binocular combination[J]. IEEE Trans Med Imaging, 2019, 38(1): 33−45. doi: 10.1109/TMI.2018.2853808
[9] Cao G, Huang L H, Tian H W, et al. Contrast enhancement of brightness-distorted images by improved adaptive gamma correction[J]. Comput Electr Eng, 2018, 66: 569−582. doi: 10.1016/j.compeleceng.2017.09.012
[10] 刘光辉, 杨琦, 孟月波, 等. 一种并行混合注意力的渐进融合图像增强方法[J]. 光电工程, 2023, 50(4): 220231. doi: 10.12086/oee.2023.220231
Liu G H, Yang Q, Yuan Y B, et al. A progressive fusion image enhancement method with parallel hybrid attention[J]. Opto-Electron Eng, 2023, 50(4): 220231. doi: 10.12086/oee.2023.220231
[11] Fan X Y, Sun Z Q, Tian E G, et al. Medical image contrast enhancement based on improved sparrow search algorithm[J]. Int J Imaging Syst Technol, 2023, 33(1): 389−402. doi: 10.1002/ima.22794
[12] An Z H, Xu C, Qian K, et al. EIEN: endoscopic image enhancement network based on retinex theory[J]. Sensors, 2022, 22(14): 5464. doi: 10.3390/s22145464
[13] 徐胜军, 杨华, 李明海, 等. 基于双频域特征聚合的低照度图像增强[J]. 光电工程, 2023, 50(12): 230225. doi: 10.12086/oee.2023.230225
Xu S J, Yang H, Li M H, et al. Low-light image enhancement based on dual-frequency domain feature aggregation[J]. Opto-Electron Eng, 2023, 50(12): 230225. doi: 10.12086/oee.2023.230225
[14] 姜鸿鹏, 章科建, 袁波, 等. 一种血管内窥镜图像增强算法[J]. 光电工程, 2019, 46(1): 180167. doi: 10.12086/oee.2019.180167
Jiang H P, Zhang K J, Yuan B, et al. A vascular enhancement algorithm for endoscope image[J]. Opto-Electron Eng, 2019, 46(1): 180167. doi: 10.12086/oee.2019.180167
[15] Hessel C, Morel J M. An extended exposure fusion and its application to single image contrast enhancement[C]// Proceedings of 2020 IEEE Winter Conference on Applications of Computer Vision, Snowmass, USA, 2020: 137–146. https://doi.org/10.1109/WACV45572.2020.9093643.
[16] Ao J, Ma C B. Adaptive stretching method for underwater image color correction[J]. Int J Pattern Recognit Artif Intell, 2018, 32(2): 1854001. doi: 10.1142/S0218001418540010
[17] Zhu Y Z, Xu C, Feng B, et al. Research on endoscopic image enhancement algorithm based on contrast fusion[C]// Proceedings of the 2021 IEEE 6th International Conference on Signal and Image Processing, Nanjing, China, 2021: 210–215. https://doi.org/10.1109/ICSIP52628.2021.9689033.
[18] Borgli H, Thambawita V, Smedsrud P H, et al. HyperKvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy[J]. Sci Data, 2020, 7(1): 283. doi: 10.1038/s41597-020-00622-y
[19] Wei C, Wang W J, Yang W H, et al. Deep retinex decomposition for low-light enhancement[C]//Proceedings of British Machine Vision Conference 2018, Newcastle, UK, 2018: 155.
[20] Guo C L, Li C Y, Guo J C, et al. Zero-reference deep curve estimation for low-light image enhancement[C]//Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, USA, 2020: 1777–1786. https://doi.org/10.1109/CVPR42600.2020.00185.
[21] Wang L F, Wu B B, Wang X, et al. Endoscopic image luminance enhancement based on the inverse square law for illuminance and retinex[J]. Int J Med Robot, 2022, 18(4): e2396. doi: 10.1002/rcs.2396
[22] Mohapatra S, Kumar Pati G, Mishra M, et al. Gastrointestinal abnormality detection and classification using empirical wavelet transform and deep convolutional neural network from endoscopic images[J]. Ain Shams Eng J, 2023, 14(4): 101942. doi: 10.1016/j.asej.2022.101942
-
点击扫一扫
图(6)
表(6)
计量
- 文章访问数:
- PDF下载数:
- 施引文献: 0
