Fluorescence Microscopy Denoising (FMD) dataset
收藏资源简介:
FMD数据集是由圣母大学创建的,专注于Poisson-Gaussian去噪的荧光显微镜图像数据集。该数据集包含12,000张真实荧光显微镜图像,这些图像来自商业共聚焦、双光子和宽场显微镜,涵盖细胞、斑马鱼和小鼠脑组织等代表性生物样本。通过图像平均技术,有效地获取了地面实况图像和60,000张不同噪声水平的噪声图像。FMD数据集用于基准测试10种代表性去噪算法,并发现深度学习方法表现最佳。该数据集是首个用于Poisson-Gaussian去噪目的的真实显微镜图像数据集,对于生物医学研究中的高质量实时去噪应用具有重要意义。
The FMD dataset, developed by the University of Notre Dame, is a fluorescence microscopy image dataset specialized for Poisson-Gaussian denoising. It comprises 12,000 real fluorescence microscopy images captured by commercial confocal, two-photon, and wide-field microscopes, covering representative biological specimens including cells, zebrafish, and mouse brain tissues. Ground truth images and 60,000 noisy images with varying noise levels were efficiently obtained via image averaging techniques. The FMD dataset was employed to benchmark 10 representative denoising algorithms, and deep learning-based methods were found to deliver the best performance. Notably, this is the first real microscopy image dataset designed specifically for Poisson-Gaussian denoising, which holds critical significance for high-quality real-time denoising applications in biomedical research.




