Deeptangle Dataset: Labelled Experimental and Synthetic Videos of Swimming and Overlapping C. elegans worms

This repository contains the dataset employed in the paper Fast spline detection in high density microscopy data. Three files are provided: 1. <em>videos.zip</em>: raw experimental videos.<br> 2. <em>labeled_data.zip</em>: labelled sections of experimental videos used for evaluation<br> 3. <em>syntehthic_dataset.zip</em>: synthetic dataset used for training <strong>Labelled data</strong> Sections are named as VIDEONAME_FRAMENUMBER_SECTIONID.<br> All labels are in labels.json and correspond to the middle frame of the clip (05.png).<br> Example plotting script is provided (<em>plot_data.py</em>). <strong>Synthetic data</strong> Clips are named as NUMBEROFWORMS_ID.<br> Labels (for all frames) are stored in labels.npy.<br> Example plotting script is provided (plot_data.py). <strong>Related</strong> Paper: https://arxiv.org/abs/2301.04460 Deeptangle code: https://github.com/kirkegaardlab/deeptangle Labelling tool: https://github.com/kirkegaardlab/deeptanglelabel --- If used, please cite <em>Albert Alonso &amp; Julius B. Kirkegaard. Fast spline detection in high density microscopy data. 2023.</em>本仓库收录了论文《高密度显微数据中的快速样条检测（Fast spline detection in high density microscopy data）》所使用的配套数据集。本次共提供3个文件： 1. *videos.zip*：原始实验视频 2. *labeled_data.zip*：用于模型评估的实验视频标注片段 3. *synthetic_dataset.zip*：用于模型训练的合成数据集 **标注数据集** 片段的命名格式为：VIDEONAME_FRAMENUMBER_SECTIONID。所有标注均存储于`labels.json`文件中，且对应片段的中间帧为`05.png`。本次附带了示例可视化脚本*plot_data.py*。 **合成数据集** 片段的命名格式为：NUMBEROFWORMS_ID。所有帧的标注均存储于`labels.npy`文件中。本次附带了示例可视化脚本*plot_data.py*。 **相关资源** 论文链接：https://arxiv.org/abs/2301.04460 Deeptangle代码仓库：https://github.com/kirkegaardlab/deeptangle 标注工具仓库：https://github.com/kirkegaardlab/deeptanglelabel 若您使用本数据集，请引用以下文献：*Albert Alonso & Julius B. Kirkegaard. Fast spline detection in high density microscopy data. 2023.*