Example dataset MiNTiF semantic segmentation

This data set serves as an example for the application of the MiNTiF plugin (https://github.com/CAiM-lab/MiNTiF). Users can test the data set creation, training, prediction, and reconstruction functionality on this data set. The pre-trained model will predict semantic segmentation masks for Arteries and Sinusoids based using the markers DAPI, Endomucin, CXCL12, Endoglin and Collagen as input. <strong>raw training data</strong> contains the original training samples to train a model. <strong>pretrained model </strong>contains a pretrained model that can be applied on MiNTiF datasets. Use MiNTiF to apply this model to "<em>tiled mintif file.h5"</em>. <strong>test data for prediction </strong>contains a dataset that can be used to test the prediction and reconstruction in MiNTiF <strong>info_channel_indices.txt </strong>contains the marker channels used in this dataset and the corresponding indices to use in MiNTiF . <strong>original_raw_file.ims</strong> is the original image file before conversion to <em>tiled mintif file.h5 </em>. Use this to test the dataset creation <strong>tiled mintif file.h5</strong> is the MiNTiF file converted from original_file.ims. Predict on this file <strong>reconstructed_file_with_prediction.tiff </strong>is the image in <em>tiled mintif file.h5 </em> reconstructed after prediction. It now contains two additional channel with the prediction channels.本数据集为MiNTiF插件（MiNTiF，https://github.com/CAiM-lab/MiNTiF）的应用示例。用户可在此数据集上测试数据集创建、模型训练、预测及图像重建功能。本预训练模型将以DAPI、Endomucin、CXCL12、Endoglin与Collagen标记物作为输入，预测动脉与血窦的语义分割掩码。 **原始训练数据集（raw training data）**：包含用于模型训练的原始训练样本。 **预训练模型（pretrained model）**：包含可应用于MiNTiF数据集的预训练模型，请使用MiNTiF将该模型应用于tiled mintif file.h5。 **预测测试数据集（test data for prediction）**：包含可用于测试MiNTiF中预测与重建功能的数据集。 **info_channel_indices.txt**：记录了本数据集使用的标记物通道及其在MiNTiF中对应的索引。 **original_raw_file.ims**：为转换为tiled mintif file.h5前的原始图像文件，可用于测试数据集创建流程。 **tiled mintif file.h5**：由original_raw_file.ims转换得到的MiNTiF格式文件，可在此文件上执行预测任务。 **reconstructed_file_with_prediction.tiff**：基于tiled mintif file.h5完成预测后重建的图像文件，新增了两个包含预测结果的通道。