Deep Learning for Microfluidic Assisted Caenorhabditis elegans Multi-parameter Identification Using YOLOv7

The effectiveness of deep learning model relies on a large number of labeled image datasets. We have collected a dataset of 3373 worm images from microfluidic devices in various studies as datasets. Then, the datasets were labeled for training methods. The acquired videos were continuously intercepted at 10-frame intervals to capture cropped worm images. Totally 3931 images were extract. Worms in each image were manually annotated using LabelImg. This VOC-format annotation tool generated Extensible Markup Language (XML) files for the model. There were 2426 labels for the WT category and 1505 for the GFP category.