how can I get these pose-stick mask pictures?

[sunshinewhy]

how can I get these pose-stick mask pictures?
Such as:

Thanks！

[AliaksandrSiarohin]

In pose_utils.py there are function draw_pose_from_cords for this purpose.

[sunshinewhy]

Thanks a lot! But I have some problems on parameter settings, could you tell me how to set them.When I run coordinates.py, it shows:
usage: compute_coordinates.py [-h] [--output_dir OUTPUT_DIR]
[--batch_size BATCH_SIZE]
[--training_ratio TRAINING_RATIO]
[--l1_penalty_weight L1_PENALTY_WEIGHT]
[--gan_penalty_weight GAN_PENALTY_WEIGHT]
[--tv_penalty_weight TV_PENALTY_WEIGHT]
[--lstruct_penalty_weight LSTRUCT_PENALTY_WEIGHT]
[--number_of_epochs NUMBER_OF_EPOCHS]
[--content_loss_layer CONTENT_LOSS_LAYER]
[--checkpoints_dir CHECKPOINTS_DIR]
[--checkpoints_ratio CHECKPOINTS_RATIO]
[--generator_checkpoint GENERATOR_CHECKPOINT]
[--discriminator_checkpoint DISCRIMINATOR_CHECKPOINT]
[--nn_loss_area_size NN_LOSS_AREA_SIZE]
[--use_validation USE_VALIDATION]
[--dataset {market,fashion,fashion128,fashion128128}]
[--display ratio DISPLAY RATIO]
[--start_epoch START_EPOCH]
[--pose_estimator POSE_ESTIMATOR]
[--images_for_test IMAGES_FOR_TEST]
[--use_input_pose USE_INPUT_POSE]
[--warp_skip {none,full,mask,stn}]
[--warp_agg {max,avg}]
[--dis_type {call,sim,warp}]
[--generated images dir GENERATED IMAGES DIR]
[--load_generated_images LOAD_GENERATED_IMAGES]
[--use_dropout_test USE_DROPOUT_TEST]
compute_coordinates.py: error: unrecognized arguments: --data_url=s3://pose-transfer/market_data/train/ --num_gpus=1

[AliaksandrSiarohin]

If I remember correctly

python compute_coordinates.py --pose_estimator pose_estimator.h5 --dataset market

In this case you should have folders with images called data/market-dataset/test and data/market-dataset/train