pix2pix_HUVEC_nuclei_immuno_cells_dataset
收藏资源简介：
This repository contains a Pix2Pix deep learning model designed to generate synthetic nuclear staining from brightfield images of circulating immune cells. The model was trained on a dataset of 226 paired brightfield and fluorescent microscopy images, which were augmented computationally by a factor of 8 to enhance model performance. The model was trained over 400 epochs using a patch size of 512x512, a batch size of 1, and a vanilla GAN loss function. The final model was selected based on quality metric scores and visual comparison to ground truth images, achieving an average SSIM score of 0.756 and an LPIPS score of 0.130. Specifications Model: Pix2Pix for generating synthetic nuclear staining from brightfield images of circulating immune cells Training Dataset: Immune Cells: 226 paired brightfield and fluorescent microscopy images Microscope: Nikon Eclipse Ti2-E, brightfield/fluorescence microscope with a 20x objective Data Type: Brightfield and fluorescent microscopy images File Format: TIFF (.tif), 16-bit Image Size: 1024 x 1022 pixels (Pixel size: 650 nm) Training Parameters: Epochs: 400 Patch Size: 512 x 512 pixels Batch Size: 1 Loss Function: Vanilla GAN loss function Model Performance: Immune Cells: SSIM Score: 0.756 LPIPS Score: 0.130 Model Selection: Models were chosen based on quality metric scores and visual inspection compared to ground truth images. Model Training: Conducted using ZeroCostDL4Mic (https://github.com/HenriquesLab/ZeroCostDL4Mic/wiki) Reference Fast label-free live imaging reveals key roles of flow dynamics and CD44-HA interaction in cancer cell arrest on endothelial monolayers Gautier Follain, Sujan Ghimire, Joanna W. Pylvänäinen, Monika Vaitkevičiūtė, Diana Wurzinger, Camilo Guzmán, James RW Conway, Michal Dibus, Sanna Oikari, Kirsi Rilla, Marko Salmi, Johanna Ivaska, Guillaume Jacquemet bioRxiv 2024.09.30.615654; doi: https://doi.org/10.1101/2024.09.30.615654
本仓库包含一款Pix2Pix深度学习模型，旨在从循环免疫细胞的明场图像生成合成的细胞核染色图像。该模型基于226组成对的明场与荧光显微图像数据集进行训练，并通过计算数据增强方式将数据集扩充8倍以提升模型性能。模型训练共进行400个训练轮次，采用512×512的图像块尺寸、批次大小为1，且使用原始生成对抗网络（vanilla GAN）损失函数。最终模型通过质量指标得分及与真实标注图像（ground truth）的视觉对比进行筛选，其平均结构相似性指数（Structural Similarity Index, SSIM）得分为0.756，学习感知图像块相似度（Learned Perceptual Image Patch Similarity, LPIPS）得分为0.130。 规格参数 模型：用于从循环免疫细胞明场图像生成合成细胞核染色的Pix2Pix模型 训练数据集： 免疫细胞相关数据集：226组成对的明场与荧光显微图像 显微镜设备：尼康Eclipse Ti2-E明场/荧光显微镜，搭配20倍物镜 数据类型：明场与荧光显微图像 文件格式：TIFF（.tif），16位 图像尺寸：1024×1022像素（像素尺寸：650 nm） 训练参数： 训练轮次（epoch）：400 图像块尺寸：512×512像素 批次大小：1 损失函数：原始生成对抗网络（vanilla GAN）损失函数 模型性能： 免疫细胞相关指标： 结构相似性指数（Structural Similarity Index, SSIM）得分：0.756 学习感知图像块相似度（Learned Perceptual Image Patch Similarity, LPIPS）得分：0.130 模型筛选：基于质量指标得分及与真实标注图像（ground truth）的视觉检查结果选择最优模型。 模型训练：基于ZeroCostDL4Mic平台完成训练（https://github.com/HenriquesLab/ZeroCostDL4Mic/wiki） 参考文献 《无标记快速活细胞成像揭示流动力学与CD44-HA相互作用在癌细胞黏附于内皮单层中的关键作用》 作者：Gautier Follain、Sujan Ghimire、Joanna W. Pylvänäinen、Monika Vaitkevičiūtė、Diana Wurzinger、Camilo Guzmán、James RW Conway、Michal Dibus、Sanna Oikari、Kirsi Rilla、Marko Salmi、Johanna Ivaska、Guillaume Jacquemet 预印本发布于bioRxiv 2024.09.30.615654；DOI：https://doi.org/10.1101/2024.09.30.615654
