sarcgraph.sg.SarcGraph.sarcomere_detection

SarcGraph.sarcomere_detection(file_path=None, raw_frames=None, segmented_zdiscs=None, tracked_zdiscs=None, sigma=1.0, min_length=8, tp_depth=4.0, full_track_ratio=0.75, skip_merging=False, c_avg_length=1, c_angle=1, c_diff_length=1, l_avg=12.0, score_threshold=0.01, angle_threshold=1.2, save_output=True)[source]

Detect sarcomeres in a video/image. The input could be the address to the video/image sample (file_path), raw frames as a numpy array (raw_frames), segmented zdiscs information in a pandas dataframe (segmented_zdiscs), or a dataframe of tracked zdiscs (tracked_zdiscs).

Parameters:

  • file_path (str) – The address of an image or a video file to be loaded

  • raw_frames (np.ndarray, shape=(frames, dim_1, dim_2, channels)) – Raw input image or video given as a 4 dimensional array

  • segmented_zdiscs (pd.DataFrame) – Information of all detected zdiscs in every frame.

  • tracked_zdiscs (pd.DataFrame) – Information of tracked zdiscs

  • sigma (float) – Standard deviation for Gaussian kernel, by default 1.0

  • min_length (int) – Minimum length for zdisc contours measured in pixels, by default 8

  • tp_depth (float, optional) – the maximum distance features can move between frames, by default 4.0

  • full_track_ratio (float, optional) – Frame ratio to seperate partially tracked vs fully tracked sarcomeres, by default 0.75

  • skip_merging (bool, optional) – skipping the merging step will result in fewer fully tracked zdiscs, by default False

  • c_avg_length (float, optional) – comparison of the length of a connection with l_avg affects the connection score, c_avg_length sets the relative effect of this score metric compared to the other two, by default 1.0

  • c_angle (float, optional) – the angle between a connection and its neighboring connections affects the connection score, c_angle sets the relative effect of this score metric compared to the other two, by default 1.0

  • c_length_diff (float, optional) – comparison of the length of a connection with the length of all neighboring connections affects the score, c_length_diff sets the relative effect of this score metric compared to the other two, by default 1.0

  • l_avg (float, optional) – an initial guess for the average length of sarcomeres in pixels, by default 15.0

  • score_threshold (float) – any connection with a score less than the threshold will be removed, by default 0.1

  • angle_threshold (float) – if a zdisc has two valid connection the angle between must be higher than the theshold, otherwise the connection with a lower score will be removed, by default 1.2 (in radians)

  • save_output (bool) – by default True

Returns:

  • pd.DataFrame – Detected sarcomeres information. Columns are 'frame' (frame number), 'sarc_id' (sarcomere id), 'zdiscs' (particle id of the two zdiscs forming a sarcomere), 'x' and 'y' (sarcomere center position), 'length', 'width', and 'angle' (sarcomere length, width, and angle).

  • List[nx.Graph] – A list of graphs each indicating connected sarcomeres (myofibrils)

Notes

See also

sarcgraph.sg.SarcGraph.zdisc_segmentation(), sarcgraph.sg.SarcGraph.zdisc_tracking()