nicetoolbox.detectors.feature_detectors.leaning.utils

Utility functions for visualizing leaning angles and calculating the angle between three points.

Functions

calculate_angle_btw_three_points

Calculate the angle between three points in a 3D space.

create_video_canvas

Create a matplotlib figure and canvas for creating a video with two subplots for PersonL and PersonR.

frame_with_linegraph

Combine a video frame with the plots for PersonL and PersonR up to the current frame.

visualize_lean_in_out_per_person

Visualize the leaning angle between midpoint of shoulders, hips, and knees for each person.
nicetoolbox.detectors.feature_detectors.leaning.utils.calculate_angle_btw_three_points(data)[source]

Calculate the angle between three points in a 3D space.

Parameters:

data (numpy.ndarray) – A 4D numpy array with shape (num_subjects, num_frames, num_points, num_coordinates) representing the coordinates of points A, B, and C for each subject and frame.

Returns:

A 4D numpy array with shape (num_subjects, num_frames,

num_points, 1) containing the angle between the three points for each subject and frame.

Return type:

numpy.ndarray

nicetoolbox.detectors.feature_detectors.leaning.utils.create_video_canvas(num_of_frames, global_min, global_max)[source]

Create a matplotlib figure and canvas for creating a video with two subplots for PersonL and PersonR.

Parameters:

  • num_of_frames (int) – The total number of frames in the video.

  • global_min (float) – The global minimum value for the y-axis.

  • global_max (float) – The global maximum value for the y-axis.

Returns:

The matplotlib figure object. canvas (matplotlib.backends.backend_agg.FigureCanvasAgg): The matplotlib canvas

object.

axL (matplotlib.axes._subplots.AxesSubplot): The left subplot axis for PersonL. axR (matplotlib.axes._subplots.AxesSubplot): The right subplot axis for PersonR.

Return type:

fig (matplotlib.figure.Figure)

nicetoolbox.detectors.feature_detectors.leaning.utils.frame_with_linegraph(frame, current_frame, data, fig, canvas, axL, axR)[source]

Combine a video frame with the plots for PersonL and PersonR up to the current frame.

Parameters:

  • frame (numpy.ndarray) – The current video frame as a numpy array.

  • current_frame (int) – The index of the current frame.

  • data (list) – A list containing two numpy arrays, dataL and dataR, representing the leaning angles for PersonL and PersonR respectively.

  • fig (matplotlib.figure.Figure) – The matplotlib figure object.

  • canvas (matplotlib.backends.backend_agg.FigureCanvasAgg) – The matplotlib canvas object.

  • axL (matplotlib.axes._subplots.AxesSubplot) – The left subplot axis for PersonL.

  • axR (matplotlib.axes._subplots.AxesSubplot) – The right subplot axis for PersonR.

Returns:

A numpy array representing the combined video frame and the plots.

Return type:

numpy.ndarray

nicetoolbox.detectors.feature_detectors.leaning.utils.visualize_lean_in_out_per_person(hip_angle, person_list, output_folder, camera_names=None)[source]

Visualize the leaning angle between midpoint of shoulders, hips, and knees for each person.

Parameters:

  • hip_angle (numpy.ndarray) – A 4D numpy array with shape (num_subjects, num_cameras, num_frames, 2) representing the leaning angles (axis angle and derivative) for each subject, camera, and frame.

  • person_list (list) – A list of strings representing the names of the persons in frame.

  • output_folder (str) – The path to the output folder where the plots will be saved.

  • camera_names (list, optional) – A list of strings representing the names of the cameras. Defaults to None.

Returns:

None