『论文笔记』CARAFE：Content-AwareReAssemblyofFEatures 『论⽂笔记』CARAFE：Content-AwareReAssemblyofFEatures 对上采样的优化：动态的选取新像素对哪些原有像素存在依赖 Large field of view. Unlike previous works (e.g. bilinear interpolation) that only exploit subpixel neighborhood, CARAFE can aggregate ...
Large field of view. Unlike previous works (e.g. bilinear interpolation) that only exploit subpixel neighborhood, CARAFE can aggregate contextual information within a large receptive field. Content-aware handling. Instead of using a fixed kernel for all samples (e.g. deconvolution), CARAFE enables...
CARAFE: Regional Airborne Greenhouse Gases Eddy Covariance Measurements, 2016-2017 简介 该数据集提供了美国宇航局 (NASA) 2016 年和 2017 年空中碳空气通量实验 (CARAFE) 期间收集的高空间分辨率空中二氧化碳、甲烷、显热和潜热涡旋协方差 (EC) 通量数据。CARAFE 使用了 NASA C-23 Sherpa 运输机,该运输机配备了...