PolyUDataset
收藏资源简介:
PolyUDataset是由香港理工大学构建的一个新的基准数据集,旨在推动真实世界噪声图像去噪的研究。该数据集包含了不同自然场景下由不同相机在不同设置下拍摄的全面真实世界噪声图像。数据集通过多次拍摄同一静态场景并计算平均图像来生成‘ground truth’图像,以更准确地评估去噪方法的性能。PolyUDataset适用于评估和开发新的真实世界图像去噪方法,特别是在解决复杂噪声问题方面。
PolyUDataset is a novel benchmark dataset constructed by The Hong Kong Polytechnic University, intended to advance research on real-world noisy image denoising. This dataset features comprehensive real-world noisy images captured by diverse cameras under varying settings across a wide range of natural scenes. The "ground truth" images are generated by repeatedly capturing the same static scene and computing the average of the captured frames, enabling more accurate performance evaluation of denoising methods. PolyUDataset is suitable for evaluating and developing novel real-world image denoising methods, especially for tackling complex noise-related problems.

- 1Real-world Noisy Image Denoising: A New Benchmark香港理工大学 · 2018年



