LINCellularNeuroscience/VAME

possible bug in code in align_egocentrical.py

WeissShahaf opened this issue · 1 comments

def alignment(path_to_file, filename, pose_ref_index, video_format, crop_size, confidence, use_video=False, check_video=False):

#read out data
data = pd.read_csv(os.path.join(path_to_file,'videos','pose_estimation',filename+'.csv'), skiprows = 2)
data_mat = pd.DataFrame.to_numpy(data)
data_mat = data_mat[:,1:] 

line 258 in align_egocentrical.py, throws out the first column (0):
data_mat = data_mat[:,1:]
so the first X coordinates in the DLC csv file are discarded.

Hi and thank you for bringing this up. This part of the code takes care of the first column, which usually is the index of the csv file, so 1, 2, 3, ...., N up to the number of frames N. But if this is for some reason not true for your data, you should adjust this part in your code.

Cheers,
Kevin