InShaPe dataset: Large-scale multi-shape phase retrieval dataset for PBF-LB/M laser beam shaping SLM phase mask estimation.

This simulated InShaPe dataset containing six beam shapes is used for training and testing on the complex forward path application, PBF-LB/M beam shaping system, in our paper "Efficient Gerchberg-Saxton algorithm deep unrolling for phase retrieval with complex forward path". In this dataset, there are 6 sub-datasets stored in 6 zip folders named by "dense+[beamshape name]+30k_pre.zip". For each beam shape, the trainging set has 10000 pairs of aberrated PBF-LB/M beam shape samples. Each paired sample contains one intensity profile of the aberrated beam shape in an imaging plane 5 times Rayleigh length before the back focal plane and one corresponding aberrated SLM phase mask that generated the aberrated beam shape. The aberrations in the paired samples are modelled with linear combination of Zernike polynomials amplified by randomly sampled Zernike coefficients.In order for the convenience of distributing the dataset, this repository does not contain the simulated paired samples. Instead, we give the original Zernike coefficients stored in Numpy array .npy files and the corresponding simulation script so that the users can run the simulation to reproduce the entire dataset. The Zernike coefficients .npy files generating the training set and test set reside in the subfolders "Output_Data" and "Val_Data", respectively, in each beam shape's zip folder. The simulation scripts generating the training set and test set reside in the root directory of each beam shape's zip folder named "data_processing_train.py" and "data_processing_test.py", respectively. Run the simulation scripts then the paired samples for training and testing will be generated in the subfolders "training_set" and "test_sample" in the zip folder of each beam shape. The generated intensity profiles and phase masks are also in numpy .npy files.

本仿真构建的InShaPe数据集包含六种光束形状，用于本文《面向复前向路径相位恢复的高效格奇贝格-桑顿算法深度展开》中针对复前向路径应用场景——PBF-LB/M光束整形系统的训练与测试。该数据集包含6个子数据集，分别存储于6个以“dense+[光束形状名称]+30k_pre.zip”命名的压缩文件夹中。针对每种光束形状，其训练集包含10000组带像差的PBF-LB/M光束形状样本对。每组样本对包含两项内容：一是位于背焦平面前方5倍瑞利长度处的成像平面内的带像差光束强度分布，二是生成该带像差光束对应的SLM（空间光调制器，Spatial Light Modulator）相位掩模。样本对中的像差通过泽尼克多项式的线性组合建模，系数由随机采样得到。为便于数据集分发，本仓库未直接存储仿真生成的样本对，而是提供了以Numpy数组格式存储的原始泽尼克系数（.npy文件）与对应仿真脚本，用户可运行脚本复现完整数据集。生成训练集与测试集的泽尼克系数.npy文件，分别存于各光束形状压缩文件夹下的“Output_Data”与“Val_Data”子文件夹中。生成训练集与测试集的仿真脚本，分别位于各光束形状压缩文件夹的根目录，命名为“data_processing_train.py”与“data_processing_test.py”。运行上述仿真脚本后，训练集与测试集样本对将分别生成于各光束形状压缩文件夹下的“training_set”与“test_sample”子文件夹中。最终生成的光束强度分布与相位掩模均以Numpy .npy文件格式存储。