Dataset: 5D imaging of freezing emulsions with solute effects

Welcome to this dataset ! <br><br>This dataset and code were used and developed for the paper "5D imaging of freezing emulsions with solute effects", by Dmytro Dedovets, Cécile Monteux, and Sylvain Deville, published in Science, April 20, 2018.<br><br>The paper is here: http://science.sciencemag.org/content/360/6386/303<br><br>The Python code has been developed to analyse the confocal images obtained during the freezing of emulsions. <br><br>If you want to use this code and the data we provided, you should organise your files as follows. Create a main directory with 2 subdirectory:<br>- "code": put the three notebooks here. Create a folder inside called "functions". In this functions folder, you need to put the detect peak script which you can find here http://nbviewer.jupyter.org/github/demotu/BMC/blob/master/notebooks/DetectPeaks.ipynb<br>- "data": put the data here<br><br>The dataset consists in image stacks (8.7Gb total). For each interface velocity, we have two sequences: one with the fluorescent channel of the liquid (filename ends with "-solute"), and a second one with the fluorescent channel of the droplet. We analyse both sequences separately in the code.<br><br>We tested the following interface velocities: 1, 2, 3, 4, 5, 6, 8, and 10 microns/s, for a temperature gradient of 5°C/mm.<br><br>When you start the notebook, another directory ("Results") will be created, and all the .csv files and figures generated will be saved there.<br><br>We are not Python experts, but the code runs relatively fast. The whole analysis for one sequence runs in a few minutes on a 2016 MacBook Pro (3,3 GHz Intel Core i7).<br><br>We still work actively on the code to improve it, so any feedback and suggestions are warmly welcome. Please keep in touch with us if you want to discuss anything.<br><br>-Sylvain Deville<br>Laboratoire de Synthèse et Fonctionnalisation des Céramiques, Cavaillon, France<br>Email: [e-mail], also on twitter @DevilleSy<br><br>

欢迎使用本数据集！<br><br>本数据集与配套代码由德米特罗·德多韦茨（Dmytro Dedovets）、塞西尔·蒙特勒（Cécile Monteux）与西尔万·德维尔（Sylvain Deville）为发表于2018年4月20日《科学》（Science）期刊的论文《溶质效应下冷冻乳液的五维成像》开发并使用。<br><br>论文链接：http://science.sciencemag.org/content/360/6386/303<br><br>本Python代码专为分析乳液冷冻过程中获取的共聚焦（confocal）图像而开发。<br><br>若需使用本代码与我们提供的数据集，请按以下方式组织文件结构：首先创建一个主目录，并在其中新建两个子目录：<br>- "code"子目录：将三个Jupyter笔记本（notebook）文件放入该目录，并在其中新建名为"functions"的文件夹，将可从http://nbviewer.jupyter.org/github/demotu/BMC/blob/master/notebooks/DetectPeaks.ipynb获取的峰值检测脚本放入此functions文件夹。<br>- "data"子目录：将数据集文件放入该目录。<br><br>本数据集为图像序列栈（image stacks），总大小8.7GB。针对每一种界面速度，均包含两组图像序列：一组为液体荧光通道图像（文件名以"-solute"结尾），另一组为液滴荧光通道图像。代码将分别对两组序列进行分析。<br><br>本次实验在5℃/mm的温度梯度下，测试了1、2、3、4、5、6、8与10 μm/s的界面速度。<br><br>运行Jupyter笔记本时，系统将自动新建"Results"目录，所有生成的.csv文件与图像均将保存至该目录。<br><br>尽管团队并非Python开发专家，但本代码运行效率较高。在2016款MacBook Pro（3,3 GHz Intel Core i7处理器）上，单组序列的完整分析仅需数分钟即可完成。<br><br>我们仍在持续优化本代码，热忱欢迎各位提出反馈与改进建议。若有任何讨论意向，欢迎随时与我们联系。<br><br>——西尔万·德维尔（Sylvain Deville）<br>法国卡瓦永陶瓷合成与功能化实验室（Laboratoire de Synthèse et Fonctionnalisation des Céramiques）<br>电子邮箱：[e-mail]，Twitter账号：@DevilleSy