A Novel Hybrid Endoscopic Dataset
收藏资源简介:
本研究介绍了一种新型混合内窥镜数据集,旨在评估基于机器学习的图像增强模型。该数据集由蒙特雷科技大学等机构创建,包含2216条合成生成的过曝光和欠曝光帧及其对应的地面实况图像。数据集通过生成对抗网络技术生成,用于探索浅层和深度学习基础的图像增强方法。此数据集主要应用于医学成像领域,特别是内窥镜检查中的图像增强,以解决由于光照条件变化导致的诊断信息缺失问题。
This study introduces a novel hybrid endoscopic dataset designed to evaluate machine learning-based image enhancement models. Developed by institutions including the Monterrey Institute of Technology and Higher Education, this dataset contains 2216 synthetic overexposed and underexposed frames along with their corresponding ground truth images. Generated using generative adversarial network (GAN) techniques, this dataset is intended to explore image enhancement methods based on both shallow and deep learning. It is primarily applied in the field of medical imaging, particularly for image enhancement in endoscopic examinations, to address the issue of missing diagnostic information caused by varying lighting conditions.




