More #DeepLabCut fun w/ - tracking a beautiful skilled reaching behavior in rats with Daniel Leventhal's lab at ! . See more videos (and the code) on our website: . Preprint here:
DeepLabCut – markerless pose estimation of user-defined features with DL for all animals, including humans By Project Github
We are hiring! Are you a full stack software engineer who wants to work on an open source project? #deeplabcut is calling... contact us (emails at ) for more information! #machinelearning #deeplearning #poseEstimation
Thank you #DeepLabCut ! (Only 18 frames for annotation)
Only pretrained on ImageNet, not on hands/pianos, etc- see paper here
For more (high res) videos of our #DeepLabCut toolbox applied to mouse reaching, mice odor trail-tracking, and drosophila laying eggs
DeepLabCut: cool tool for motion tracking mice ⁦
Take DeepLabCut from and , for example: . It works by training on one task, then transferring that to the pose estimation task. Without the initial training it would require far more data points to estimate pose.
Freely available on GitHub! See for paper and code! Thanks for checking it out!
Learning a bit about “Software 2.0” — integration of annotation, training, and inference using deep learning. DeepLabCut is the tool and it’s pretty cool: #CMUOpenScience