Image-based Data on Strain Fields of Microstructures with Porosity Defects_3

This dataset is one of the three datasets that are utilized in the research, 'Predicting elastic strain fields in defective microstructures using image colorization algorithms' submitted for publication in the Computational Material Science Journal. This is a part of supplementary data for the Data in Brief article co-submitted alongside the aforementioned paper in Computational Material Science Journal. The data file is a MATLAB Binary Data Compressed file (.mat) that contains images of microstructures and their respective strain fields represented as two-dimensional arrays. This data is utilized for training a convolutional neural network to predict strain fields within a microstructure, through image colorization algorithm. This dataset contains six sub-folders each having two datasets (500 samples and 50 samples) of microstructures with porosity defects of six unique shapes described in the aforementioned research article along with their strain fields.本数据集为投稿至《计算材料科学期刊》(Computational Material Science Journal)的研究论文《利用图像着色算法预测缺陷微结构中的弹性应变场》所使用的三套数据集之一。本数据集同时也是与上述论文同期投稿至该期刊的《数据简报》(Data in Brief)配套文章的补充数据的一部分。该数据文件为MATLAB二进制压缩数据文件（.mat），其中包含以二维数组形式存储的微结构图像及其对应的应变场。本数据集通过图像着色算法，用于训练卷积神经网络（Convolutional Neural Network）以预测微结构内部的应变场。本数据集包含6个子文件夹，每个子文件夹均包含两套数据集（分别含500个样本与50个样本），对应上述研究论文中提及的6种独特孔隙缺陷形状的微结构及其应变场。