WormSwin: C. elegans Video Datasets

<p>Data used for our paper "WormSwin: Instance Segmentation of C. elegans using Vision Transformer".<br>This publication is divided into three parts:</p> <ol> <li>CSB-1 Dataset</li> <li>Synthetic Images Dataset</li> <li>MD Dataset</li> </ol> <p>The CSB-1 Dataset&nbsp;consists of frames extracted from videos of Caenorhabditis elegans (C. elegans) annotated with binary masks. Each C. elegans is separately annotated, providing accurate annotations even for overlapping instances. All annotations are provided in binary mask format and as COCO Annotation JSON files (see&nbsp;<a href="https://cocodataset.org/#format-data">COCO website</a>).</p> <p>The videos are named after the following pattern:</p> <pre><code>&lt;"worm age in hours"_"mutation"_"irradiated (binary)"_"video index (zero based)"&gt;</code></pre> <p>For mutation the following values are possible:&nbsp;</p> <ol> <li>wild type</li> <li>csb-1 mutant</li> <li>csb-1 with rescue mutation</li> </ol> <p>An example video name would be <em>24_1_1_2</em> meaning it shows C. elegans with csb-1&nbsp;mutation, being 24h old which got irradiated.</p> <p>Video data was provided by M. Rieckher; Instance Segmentation Annotations were created under supervision of K. Bozek and M. Deserno.<br><br>The Synthetic Images Dataset was created by cutting out C. elegans (foreground objects)&nbsp;from the CSB-1 Dataset&nbsp;and placing them randomly on background images also taken from the CSB-1 Dataset. Foreground objects were flipped, rotated and slightly blurred before placed on the background images.<br>The same was done with the binary mask annotations taken from CSB-1 Dataset so that they match the foreground objects in the synthetic images.&nbsp;Additionally, we added rings of random color, size, thickness and position&nbsp;to the background images to simulate petri-dish edges.</p> <p>This synthetic dataset was generated by M. Deserno.<br><br>The Mating Dataset (MD) consists of 450 grayscale image patches&nbsp;of 1,012 x 1,012 px showing C. elegans with high overlap, crawling on a petri-dish.<br>We took the patches from a 10 min. long video of size&nbsp;3,036 x 3,036 px. The video was downsampled from 25 fps to 5 fps before selecting 50 random frames for annotating and patching.<br>Like the other datasets, worms were annotated with binary masks and annotations are provided as COCO Annotation JSON files.</p> <p>The video data was provided by&nbsp;X.-L. Chu;&nbsp;Instance Segmentation Annotations were created under supervision of K. Bozek and M. Deserno.</p> <p><br>Further details about the datasets can be found in our <a href="https://doi.org/10.1038/s41598-023-38213-7" target="_blank" rel="noopener">paper</a>.</p>