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

The Caenorhabditis elegans (C. elegans) is an ideal model organism for studying human diseases and genetics due to its transparency suitability optical imaging. However, manually sorting a large population of C. experiments tedious inefficient. microfluidic-assisted chip considered promising platform address this issue automation ease operation. Nevertheless, automated with multiple parameters requires efficient identification technology the different research demands worm phenotypes. To improve efficiency accuracy multi-parameter sorting, we developed deep learning using You Only Look Once (YOLO)v7 detect recognize automatically. We used dataset 3931 annotated worms in microfluidic chips from various studies. Our showed higher precision than YOLOv5 Faster R-CNN, achieving mean average (mAP) at 0.5 intersection over union ([email protected]) threshold 99.56%. Additionally, our demonstrated good generalization ability, [email protected] 94.21% on external validation set. can efficiently accurately identify calculate phenotypes worms, including size, movement speed, fluorescence. potentially promote development integrated platforms.