Hi,

I just had two questions: (1) Can I use motionmapper to map multiple csvs from DLC to one single postural map? Do I just put in multiple file names into deeplabcut_into_motionMapper.m for the variable deeplabcutfile (e.g. deeplabcutfile = 'data1.csv','data2.csv')? Will this do what I described above?

(2) How do I map one dataset's postural time series data onto another dataset's? This was done in the original paper (section 4.4 of Berman et al.) for male and female flies.

Thanks,
Annie

Greetings

We have a trained 3D DLC system in which we want to analyze the speed and response time of short macaque monkey videos.
I went through the steps and got 3D trained network of DLC and data of the videos through the DLC.
First how do I know how much time is between two frames(we have the Excel table for a video that the DLC analyzed).
Second, I would love to know if there is an existing code that I can use to get the total length of movement for a particular limb for 1 video (e.g. a finger)

Many thanks in advance
Dvir

Hi there. I am using the time_in_ROI.py to analyze open field behavior of rats with DLC markers. Specifically, i am looking at the time spent in the center of the open field. My only ROI is the "center" box. When i run the code that outputs the time spent in the ROIs, i get odd number for the [avg_time_in_roi_sec] variable. The numbers for the time spent in the center are unrealistically high and do not match the animals movement. I attached a plot of the animal's "back" marker distance from the center. You can see that the [avg_time_in_roi_sec] variable outputs about 145 seconds, while the animal barely enters the center (the "center" box is about 80 pixels in either direction of the center). All other output variables from time_in_ROI.py match the animals behavior. I appreciate the help! - Victoria

Hi,
The Readme references pupilDLC.py. But it seems to be missing? Or am I missing something?

Thanks in advance
Balaji

Hey, thanks for DLC and DLC utils they're amazing :) I'm trying to run this piece of code for ROI analysis on Colab:

bodyparts=Dataframe.columns.get_level_values(1) #you can read out the header to get body part names!
bodyparts2plot=bodyparts #you could also take a subset, i.e. =['snout']
%matplotlib inline
PlottingResults(Dataframe,bodyparts2plot,alphavalue=.2,pcutoff=.5,fs=(8,4))

But I continue to get this "KeyError: 'likelihood'". Can you help me, please? Thanks in advance!

I have been following all instructions, but I still got the following error:

Error in deeplabcut_into_motionMapper (line 119)
[embeddingValues{i},~] = run_tSne(data,parameters);

Could you help me out with this?

Hello! I've been following the code, step by step, and it has been working up to this point:

After that, I didn't see anything (X or Y coordinates or anything related to the video)? What should I be doing now?

when I try execult
video ='IIR1DIS75T1.mp4'
DLCscorer= 'IIR1DIS75T1DLC_resnet50_APNov2shuffle1_190000_filtered'
dataname = str(Path(video).stem) + DLCscorer + '.h5'

#loading output of DLC
Dataframe = pd.read_hdf(os.path.join(dataname))

this erro is show

FileNotFoundError Traceback (most recent call last)
in
4
5 #loading output of DLC
----> 6 Dataframe = pd.read_hdf(os.path.join(dataname))

/usr/local/lib/python3.7/dist-packages/pandas/io/pytables.py in read_hdf(path_or_buf, key, mode, errors, where, start, stop, columns, iterator, chunksize, **kwargs)
425
426 if not exists:
--> 427 raise FileNotFoundError(f"File {path_or_buf} does not exist")
428
429 store = HDFStore(path_or_buf, mode=mode, errors=errors, **kwargs)

FileNotFoundError: File IIR1DIS75T1IIR1DIS75T1DLC_resnet50_APNov2shuffle1_190000_filtered.h5 does not exist

I ran the deeplabcut_into_motionMapper.m using data that I got from DLC and for some reason it stops once it generates figure 7 (see picture below). I have tried running this several times and it always pauses there. It also says "Parallel pool using he 'local' profile is shutting down."

As reported by a user, time_in_each_roi consider's the tracked point to be in the closest ROI even if it is, in fact, outside the specified roi box.
The desired behaviour is that only the time spent when the point is inside the box should be counted.

I'm reporting this here for reference and will fix it ASAP, will post when fixed.

Im running the DLCUtils function on the cluster and i get the following error:

import deeplabcut
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/deeplabcut/init.py", line 33, in
from deeplabcut import pose_estimation_tensorflow
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/deeplabcut/pose_estimation_tensorflow/init.py", line 17, in
from deeplabcut.pose_estimation_tensorflow.nnet import *
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/deeplabcut/pose_estimation_tensorflow/nnet/init.py", line 14, in
from deeplabcut.pose_estimation_tensorflow.nnet.losses import *
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/deeplabcut/pose_estimation_tensorflow/nnet/losses.py", line 5, in
import tensorflow as tf
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/tensorflow/init.py", line 22, in
from tensorflow.python import pywrap_tensorflow # pylint: disable=unused-import
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/tensorflow/python/init.py", line 49, in
from tensorflow.python import pywrap_tensorflow
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 74, in
raise ImportError(msg)
ImportError: Traceback (most recent call last):
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in
from tensorflow.python.pywrap_tensorflow_internal import *
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in
_pywrap_tensorflow_internal = swig_import_helper()
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
_mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/imp.py", line 242, in load_module
return load_dynamic(name, filename, file)
File "/icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/imp.py", line 342, in load_dynamic
return _load(spec)
ImportError: /usr/lib64/libstdc++.so.6: version `GLIBCXX_3.4.17' not found (required by /icm/hydra/home/grid/plgmirandeitor/.pyenv/versions/3.6.0/lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so)

Any idea what to do to solve the problem?

I have set the rois dictionary with 'middle' as a key and a named tuple value as follows.

position = namedtuple('position', ['topleft', 'bottomright'])
rois = {'middle': position((300, 400), (500, 800))}

bp_data.shape is (23188, 2); data from the DLC2 where 23188 frames contain one body part label.

When I executed the 'get_roi_at_each_frame(bp_data, rois)' function, I get the following error.

ValueError Traceback (most recent call last)
in ()
----> 1 get_roi_at_each_frame(bp_data, rois)

in get_roi_at_each_frame(bp_data, rois)
57 for idx, center in enumerate(centers):
58 cnt = np.tile(center, data_length).reshape((data_length, 2))
---> 59 dist = np.hypot(np.subtract(cnt[:, 0], bp_data[:, 0]), np.subtract(cnt[:, 1], bp_data[:, 1]))
60 distances[:, idx] = dist
61

ValueError: operands could not be broadcast together with shapes (2,) (23188,)

Any help would be greatly appreciated.

Hello,
I am trying to use the deeplabcut_into_motionMapper.m function but got the following error:

Mean value of sigma: 0.010111
Minimum value of sigma: 0.0022873
Maximum value of sigma: 0.087411
Iteration 1: error is 99.4207, change is 1
Error using scatter (line 110)
CData must be an RGB triplet, an M-by-1 vector of the same length as X, or an M-by-3 matrix.

Error in tsne_p_sparse (line 149)
scatter(ydata(:,1),ydata(:,2),[],parameters.signalLabels,'filled')

Error in tsne_d (line 60)
[yData,errors] = tsne_p_sparse(P, parameters, no_dims, relTol);

Error in run_tSne (line 39)
[yData,betas,P,errors] = tsne_d(D,parameters);

Error in deeplabcut_into_motionMapper (line 106)
[embeddingValues{i},~] = run_tSne(BRAC36002e,parameters);

any idea about why this happened?

I am looking to calculate food cup behavior for a rat -- specifically how much time their head spends in a recessed food cup in a wall. I see that there are currently 2 utils to calculate time spent in a region of interest: time_in_ROI.py, and the polarbean DLC_ROI tool. And I have some basic questions that I have been thinking about...

• which one of the two would be better suited for my specific task?
• if I am only looking to quantify the food cup behavior, do I still need to do certain DLC functions after training/evaluating the network such as: analyze videos, plot trajectories, filter predictions, etc? I know that I should probably create the labeled videos, but I am not sure about the others.
• I have been going through this page for time_spent_in_ROI.py and trying to understand the code. For import time_in_each_roi it says that "the function needs to be in the same folder as the notebook". What is the jupyter notebook? Is it required to use the ROI analysis? I am running DLC through my university's linux cluster so I am not sure how to incorporate the notebook into that. -- do I need to run this through a GPU or do I just type all the code out in terminal using ipython?

Thanks in advance.

This can happen when loading a DLC pickle file across systems with different python versions.

You can convert it:

!pip3 install pickle5
path_to_protocol5 = '/Users/alex/Code/dlc_playingdata/MultiMouse-Daniel-2019-12-16/training-datasets/iteration-1/UnaugmentedDataSet_MultiMouseDec16/Documentation_data-MultiMouse_95shuffle0.pickle'

import pickle5 as p
import pickle


with open(path_to_protocol5, "rb") as fh:
data = p.load(fh)

with open(path_to_protocol5, "wb") as f:
  # Pickle the 'labeled-data' dictionary using the highest protocol available.
  pickle.dump(data, f, pickle.HIGHEST_PROTOCOL)

See https://stackoverflow.com/questions/63329657/python-3-7-error-unsupported-pickle-protocol-5