Optical MNIST classification with MMF - transmission matrix + MMF speckle output

This dataset is linked to the paper from Ancora et al., "Low-power multi-mode fiber projector overcomes shallow neural networks classifiers" 2024.It includes the following data:Pre-characterized <b>transmission matrix</b> of the optical fiber (transmission_fullRes.npy)Another transmission matrix example of the entire fiber facet (from another dataset, TM00_10000_P28_handwritten_inA_cam_Alpha)<b>MNIST dataset</b> used as input for the optical fiberCorresponding <b>labels</b><b>Speckle patterns</b> recorded by the camera at the output edge of the fiber<br>

本数据集关联Ancora等人2024年发表的题为《低功率多模光纤投影仪突破浅层神经网络分类器局限》（原英文标题：Low-power multi-mode fiber projector overcomes shallow neural networks classifiers）的论文。本数据集包含以下数据：光纤的预表征<b>传输矩阵(transmission matrix)</b>（存储文件为transmission_fullRes.npy）；另一组针对整个光纤端面的传输矩阵示例（源自另一数据集TM00_10000_P28_handwritten_inA_cam_Alpha）；<b>MNIST数据集(MNIST dataset)</b>，用作该光纤的输入数据；对应的<b>标签</b>；由相机在光纤输出端面采集到的<b>散斑图样(Speckle patterns)</b><br>