Kinematics Analysis#
Motion capture is widely used in biomechanics and motor neuroscience.
DLC2Kinematics#
One way to perform kinematics analysis based on DeepLabCut outputs, is using DLC2Kinematics.
It has a simple interface and you can visualize as well as manipulate pose estimation data. Here is a simple code example:
# Importing the library:
import dlc2kinematics
# Load data data (from your predictions)
df, bodyparts, scorer = dlc2kinematics.load_data(<path of the h5 file>)
# Computing velocities:
df_vel = dlc2kinematics.compute_velocity(df,bodyparts=['all'])
# Compute acceleration for some body parts:
df_vel = dlc2kinematics.compute_acceleration(df,bodyparts=['nose','joystick'])
For more details and options, please refer to DLC2Kinematics demo notebook.
Another option is using DOKA, which Johanna Schultz mentioned and created together with Fabian Plum and others.