some parameter's meaning in the code

[zq1335030905]

I want to know what is unit meaning in this function. Hope for your reply.

`def preprocess_test(motion, meanpose, stdpose, unit=128): motion = motion * unit

motion[1, :, :] = (motion[2, :, :] + motion[5, :, :]) / 2
motion[8, :, :] = (motion[9, :, :] + motion[12, :, :]) / 2

start = motion[8, :, 0]
motion = localize_motion(motion)
motion = normalize_motion(motion, meanpose, stdpose)

return motion, start

`

[RahhulDd]

@zq1335030905 Hi, I am not the author of this repo but I have been working on something similar by following this repo. The unit parameter from what I have understood till far is the over all scaling of the skeleton image to make it smaller or bigger.

[zq1335030905]

@RahhulDd Thanks a lot. And I want to know why retargeting work needs the meanpose and stdpose and how to get it? I want to follow this repo to do experiments on my dataset. but I notice that the solo-dance dataset and mixamo dataset use the same meanpose and stdpose. Thanks a lot if you can answer it.

[RahhulDd]

@zq1335030905 The meanpose and stdpose are used to whiten the data which help in normalising each frame pose according to mean and std of all the frames in the dataset. This has been a standard practice not only for retargeting but other domains as well. This is done by taking the mean and std pose for all the motion files produced after preprocessing (This file helps in segmentation of each motion video to fixed length). This will give you a 15 x 2 array for both mean and std deviation where 15 is the number of joints and 2 is axes of projection(x,z) most probably.

[zq1335030905]

Thanks. @RahhulDd By the way, the meanpose and stdpose are for whole dataset or just for one person ?

[RahhulDd]

@zq1335030905 It is across whole data set if you stack up all the clips into one array (Say X).