IISA-DB
收藏资源简介:
IISA-DB数据集是由索尼AI和佛罗伦萨大学等机构创建的,包含785对图像和其内在尺度(IIS)的标注。该数据集通过专家在严格控制的众包研究中对图像进行标注得到,旨在评估图像在不同尺度下的感知质量,并找到图像展示最高感知质量的最大尺度。数据集可应用于图像显示和存储优化,提高质量测量灵敏度,以及创建更好的低级视觉数据集等方面。
The IISA-DB dataset was created by Sony AI, the University of Florence and other institutions, containing 785 pairs of images and their intrinsic scale (IIS) annotations. This dataset is generated via expert annotation of images in strictly controlled crowdsourcing studies, aiming to evaluate the perceived quality of images across different scales and identify the maximum scale at which images exhibit the highest perceived quality. The dataset can be applied to image display and storage optimization, improving the sensitivity of quality measurement, and creating better low-level vision datasets, among others.

- 1Image Intrinsic Scale Assessment: Bridging the Gap Between Quality and Resolution索尼AI · 2025年



