PolarNS 和 PolarBurstSR
收藏资源简介:
PolarNS是首个大规模偏振噪声统计数据集,提供了偏振图像噪声传播的详细特征。它包含54个在暗室内的物体和190个室内外场景,通过静态位置捕获的爆发图像计算得出每个场景的平均值和方差。PolarBurstSR是用于训练和评估偏振成像爆发式超分辨率模型的数据集,提供了针对偏振数据的高质量图像序列。这两个数据集旨在解决偏振成像中的噪声问题,特别是在低光照效率和低空间分辨率导致的噪声增加和偏振测量受损的场景中。
PolarNS is the first large-scale statistical dataset for polarization noise, providing detailed characteristics of noise propagation in polarization images. It encompasses 54 objects captured in a darkroom and 190 indoor and outdoor scenes, with the mean and variance of each scene calculated from burst images taken at static positions. PolarBurstSR is a dataset for training and evaluating polarization imaging burst super-resolution models, offering high-quality image sequences tailored for polarization data. These two datasets aim to address noise issues in polarization imaging, particularly in scenarios where noise is amplified and polarization measurements are degraded due to low illumination efficiency and low spatial resolution.

- 1Benchmarking Burst Super-Resolution for Polarization Images: Noise Dataset and Analysis韩国科学技术院(KAIST) · 2025年



