2023-07-19 17:37:20 +08:00
###############################################################################
## PROJECT PARAMETERS ##
###############################################################################
2023-12-08 00:08:34 +08:00
# Configure your project parameters here.
#
# IMPORTANT:
# If a parameter is not found here, Pose2Sim will look for its value in the
# Config.toml file of the level above. This way, you can set global
# instructions for the Session and alter them for specific Participants or Trials.
2024-01-05 11:20:12 +08:00
#
# If you wish to overwrite a parameter for a specific trial or participant,
2023-12-08 00:08:34 +08:00
# edit its Config.toml file by uncommenting its key (e.g., [project])
# and editing its value (e.g., frame_range = [10,300]). Or else, uncomment
# [filtering.butterworth] and set cut_off_frequency = 10, etc.
2023-07-19 17:37:20 +08:00
2023-12-06 16:48:11 +08:00
[ project ]
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
multi_person = true # true for trials with multiple participants. If false, only the main person in scene is analyzed (and it run much faster).
participant_height = [ 1.72 , 1.40 ] # m # float if single person, list of float if multi-person (same order as the Static trials) # Only used for marker augmentation
participant_mass = [ 70.0 , 63.5 ] # kg # Only used for marker augmentation and scaling
frame_rate = 'auto' # fps # int or 'auto'. If 'auto', finds from video (or defaults to 60 fps if you work with images)
frame_range = [ ] # For example [10,300], or [] for all frames.
## If cameras are not synchronized, designates the frame range of the camera with the shortest recording time
2023-12-08 00:08:34 +08:00
## N.B.: If you want a time range instead, use frame_range = time_range * frame_rate
## For example if you want to analyze from 0.1 to 2 seconds with a 60 fps frame rate,
## frame_range = [0.1, 2.0]*frame_rate = [6, 120]
exclude_from_batch = [ ] # List of trials to be excluded from batch analysis, ['<participant_dir/trial_dir>', 'etc'].
# e.g. ['S00_P00_Participant/S00_P00_T00_StaticTrial', 'S00_P00_Participant/S00_P00_T01_BalancingTrial']
2024-01-05 11:20:12 +08:00
2024-03-16 22:49:40 +08:00
[ pose ]
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
vid_img_extension = 'mp4' # any video or image extension
pose_model = 'HALPE_26' #With RTMLib: HALPE_26 (default, body and feet), COCO_133 (body, feet, hands), COCO_17 (body)
# /!\ Only RTMPose is natively embeded in Pose2Sim. For all other pose estimation methods, you will have to run them yourself, and then refer to the documentation to convert the files if needed
#With MMPose: HALPE_26 (default), COCO_133, COCO_17, CUSTOM. See example at the end of the file
#With openpose: BODY_25B, BODY_25, BODY_135, COCO, MPII
#With mediapipe: BLAZEPOSE
#With alphapose: HALPE_26, HALPE_68, HALPE_136, COCO_133
#With deeplabcut: CUSTOM. See example at the end of the file
mode = 'balanced' # 'lightweight', 'balanced', 'performance'
det_frequency = 1 # Run person detection only every N frames, and inbetween track previously detected bounding boxes ( (still run keypoint detection on all frame).
#Equal to or greater than 1, can be as high as you want in simple uncrowded cases. Much faster, but might be less accurate.
tracking = true # Gives consistent person ID across frames. Slightly slower but might facilitate synchronization if other people are in the background
display_detection = true
save_video = 'to_video' # 'to_video' or 'to_images', 'none', or ['to_video', 'to_images']
output_format = 'openpose' # 'openpose', 'mmpose', 'deeplabcut', 'none' or a list of them # /!\ only 'openpose' is supported for now
2024-03-16 22:49:40 +08:00
[ synchronization ]
2024-04-14 00:42:30 +08:00
display_sync_plots = true # true or false (lowercase)
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
keypoints_to_consider = 'all' # 'all' if all points should be considered, for example if the participant did not perform any particicular sharp movement. In this case, the capture needs to be 5-10 seconds long at least
# ['RWrist', 'RElbow'] list of keypoint names if you want to specify the keypoints to consider.
2024-04-12 07:59:59 +08:00
approx_time_maxspeed = 'auto' # 'auto' if you want to consider the whole capture (default, slower if long sequences)
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
# [10.0, 2.0, 8.0, 11.0] list of times in seconds, one value per camera if you want to specify the approximate time of a clear vertical event by one person standing alone in the scene
time_range_around_maxspeed = 2.0 # Search for best correlation in the range [approx_time_maxspeed - time_range_around_maxspeed, approx_time_maxspeed + time_range_around_maxspeed]
likelihood_threshold = 0.4 # Keypoints whose likelihood is below likelihood_threshold are filtered out
filter_cutoff = 6 # time series are smoothed to get coherent time-lagged correlation
filter_order = 4
2024-03-16 22:49:40 +08:00
2024-04-15 20:12:14 +08:00
# Take heart, calibration is not that complicated once you get the hang of it!
2023-07-19 17:37:20 +08:00
[ calibration ]
2024-02-05 18:55:21 +08:00
calibration_type = 'convert' # 'convert' or 'calculate'
2023-07-19 17:37:20 +08:00
[ calibration . convert ]
2023-12-08 00:08:34 +08:00
convert_from = 'qualisys' # 'qualisys', 'optitrack', vicon', 'opencap', 'easymocap', 'biocv', 'anipose', or 'freemocap'
2023-07-19 17:37:20 +08:00
[ calibration . convert . qualisys ]
binning_factor = 1 # Usually 1, except when filming in 540p where it usually is 2
2023-09-20 20:39:40 +08:00
[ calibration . convert . optitrack ] # See readme for instructions
2023-12-06 16:48:11 +08:00
[ calibration . convert . vicon ] # No parameter needed
[ calibration . convert . opencap ] # No parameter needed
[ calibration . convert . easymocap ] # No parameter needed
[ calibration . convert . biocv ] # No parameter needed
[ calibration . convert . anipose ] # No parameter needed
[ calibration . convert . freemocap ] # No parameter needed
2023-08-19 14:59:34 +08:00
2023-07-19 17:37:20 +08:00
2023-08-19 14:59:34 +08:00
[ calibration . calculate ]
2023-12-06 16:48:11 +08:00
# Camera properties, theoretically need to be calculated only once in a camera lifetime
[ calibration . calculate . intrinsics ]
2024-04-15 22:03:44 +08:00
overwrite_intrinsics = false # set to false if you don't want to recalculate intrinsic parameters
2023-08-19 14:59:34 +08:00
show_detection_intrinsics = true # true or false (lowercase)
intrinsics_extension = 'jpg' # any video or image extension
2023-08-21 03:55:13 +08:00
extract_every_N_sec = 1 # if video, extract frames every N seconds (can be <1 )
2023-08-19 14:59:34 +08:00
intrinsics_corners_nb = [ 4 , 7 ]
intrinsics_square_size = 60 # mm
2023-12-06 16:48:11 +08:00
# Camera placements, need to be done before every session
[ calibration . calculate . extrinsics ]
2023-12-09 19:53:43 +08:00
calculate_extrinsics = true # true or false (lowercase)
2023-12-06 16:48:11 +08:00
extrinsics_method = 'scene' # 'board', 'scene', 'keypoints'
# 'board' should be large enough to be detected when laid on the floor. Not recommended.
# 'scene' involves manually clicking any point of know coordinates on scene. Usually more accurate if points are spread out.
# 'keypoints' uses automatic pose estimation of a person freely walking and waving arms in the scene. Slighlty less accurate, requires synchronized cameras.
moving_cameras = false # Not implemented yet
2023-07-19 17:37:20 +08:00
2023-12-06 16:48:11 +08:00
[ calibration . calculate . extrinsics . board ]
show_reprojection_error = true # true or false (lowercase)
extrinsics_extension = 'png' # any video or image extension
extrinsics_corners_nb = [ 4 , 7 ] # [H,W] rather than [w,h]
extrinsics_square_size = 60 # mm # [h,w] if square is actually a rectangle
[ calibration . calculate . extrinsics . scene ]
show_reprojection_error = true # true or false (lowercase)
extrinsics_extension = 'png' # any video or image extension
# list of 3D coordinates to be manually labelled on images. Can also be a 2 dimensional plane.
# in m -> unlike for intrinsics, NOT in mm!
object_coords_3d = [ [ -2.0 , 0.3 , 0.0 ] ,
[ -2.0 , 0.0 , 0.0 ] ,
[ -2.0 , 0.0 , 0.05 ] ,
[ -2.0 , -0.3 , 0.0 ] ,
[ 0.0 , 0.3 , 0.0 ] ,
[ 0.0 , 0.0 , 0.0 ] ,
[ 0.0 , 0.0 , 0.05 ] ,
[ 0.0 , -0.3 , 0.0 ] ]
[ calibration . calculate . extrinsics . keypoints ]
# Coming soon!
2023-07-19 17:37:20 +08:00
[ personAssociation ]
2024-04-15 15:33:55 +08:00
likelihood_threshold_association = 0.3
[ personAssociation . single_person ]
reproj_error_threshold_association = 20 # px
tracked_keypoint = 'Neck' # If the neck is not detected by the pose_model, check skeleton.py
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
# and choose a stable point for tracking the person of interest (e.g., 'right_shoulder' or 'RShoulder')
2024-04-15 15:33:55 +08:00
[ personAssociation . multi_person ]
reconstruction_error_threshold = 0.1 # 0.1 = 10 cm
min_affinity = 0.2 # affinity below which a correspondence is ignored
2023-07-19 17:37:20 +08:00
[ triangulation ]
2024-02-28 23:28:09 +08:00
reorder_trc = false # only checked if multi_person analysis
2023-07-19 17:37:20 +08:00
reproj_error_threshold_triangulation = 15 # px
2024-03-06 21:11:56 +08:00
likelihood_threshold_triangulation = 0.3
2023-07-19 17:37:20 +08:00
min_cameras_for_triangulation = 2
interpolation = 'cubic' #linear, slinear, quadratic, cubic, or none
2024-01-05 21:32:55 +08:00
# 'none' if you don't want to interpolate missing points
2023-08-21 03:55:13 +08:00
interp_if_gap_smaller_than = 10 # do not interpolate bigger gaps
show_interp_indices = true # true or false (lowercase). For each keypoint, return the frames that need to be interpolated
2024-01-15 18:51:02 +08:00
handle_LR_swap = false # Better if few cameras (eg less than 4) with risk of limb swapping (eg camera facing sagittal plane), otherwise slightly less accurate and slower
2024-01-08 20:01:39 +08:00
undistort_points = false # Better if distorted image (parallel lines curvy on the edge or at least one param > 10^-2), but unnecessary (and slightly slower) if distortions are low
2024-04-16 08:28:46 +08:00
make_c3d = false # save triangulated data in c3d format in addition to trc
2023-07-19 17:37:20 +08:00
[ filtering ]
2023-08-19 14:59:34 +08:00
type = 'butterworth' # butterworth, kalman, gaussian, LOESS, median, butterworth_on_speed
2024-02-28 23:28:09 +08:00
display_figures = false # true or false (lowercase)
2024-04-16 08:28:46 +08:00
make_c3d = false # also save triangulated data in c3d format
2023-07-19 17:37:20 +08:00
[ filtering . butterworth ]
order = 4
cut_off_frequency = 6 # Hz
2023-08-19 14:59:34 +08:00
[ filtering . kalman ]
# How much more do you trust triangulation results (measurements), than previous data (process assuming constant acceleration)?
trust_ratio = 100 # = measurement_trust/process_trust ~= process_noise/measurement_noise
smooth = true # should be true, unless you need real-time filtering
2023-07-19 17:37:20 +08:00
[ filtering . butterworth_on_speed ]
order = 4
cut_off_frequency = 10 # Hz
[ filtering . gaussian ]
sigma_kernel = 2 #px
[ filtering . LOESS ]
nb_values_used = 30 # = fraction of data used * nb frames
[ filtering . median ]
kernel_size = 9
2023-12-08 00:08:34 +08:00
2024-01-20 03:03:35 +08:00
[ markerAugmentation ]
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
## Requires the following markers: ["Neck", "RShoulder", "LShoulder", "RHip", "LHip", "RKnee", "LKnee",
## "RAnkle", "LAnkle", "RHeel", "LHeel", "RSmallToe", "LSmallToe",
## "RBigToe", "LBigToe", "RElbow", "LElbow", "RWrist", "LWrist"]
2024-04-16 23:28:37 +08:00
make_c3d = false # save triangulated data in c3d format in addition to trc
2024-01-20 03:03:35 +08:00
2023-07-19 17:37:20 +08:00
[ opensim ]
2023-12-08 00:08:34 +08:00
static_trial = [ 'S00_P00_Participant/S00_P00_T00_StaticTrial' ]
2023-12-09 19:53:43 +08:00
# # If this Config.toml file is at the Trial level, set to true or false (lowercase);
2023-12-08 00:08:34 +08:00
# # At the Participant level, specify the name of the static trial folder name, e.g. ['S00_P00_T00_StaticTrial'];
# # At the Session level, add participant subdirectory, e.g. ['S00_P00_Participant/S00_P00_T00_StaticTrial', 'S00_P01_Participant/S00_P00_T00_StaticTrial']
2023-12-09 19:53:43 +08:00
opensim_bin_path = 'C:\OpenSim 4.4\bin'
2023-12-06 16:48:11 +08:00
2023-11-01 22:56:37 +08:00
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
# CUSTOM skeleton, if you trained your own model from DeepLabCut or MMPose for example.
2023-11-01 23:24:45 +08:00
# Make sure the node ids correspond to the column numbers of the 2D pose file, starting from zero.
2024-01-08 20:01:39 +08:00
#
2023-12-08 00:08:34 +08:00
# If you want to perform inverse kinematics, you will also need to create an OpenSim model
# and add to its markerset the location where you expect the triangulated keypoints to be detected.
2024-01-08 20:01:39 +08:00
#
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
# In this example, CUSTOM reproduces the HALPE_26 skeleton (default skeletons are stored in skeletons.py).
2023-12-08 00:08:34 +08:00
# You can create as many custom skeletons as you want, just add them further down and rename them.
2024-01-08 20:01:39 +08:00
#
# Check your model hierarchy with: for pre, _, node in RenderTree(model):
# print(f'{pre}{node.name} id={node.id}')
2023-11-01 22:56:37 +08:00
[ pose . CUSTOM ]
name = "CHip"
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
id = "19"
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children ] ]
name = "RHip"
2024-03-06 21:11:56 +08:00
id = 12
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children ] ]
name = "RKnee"
2024-03-06 21:11:56 +08:00
id = 14
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children ] ]
name = "RAnkle"
2024-03-06 21:11:56 +08:00
id = 16
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children . children ] ]
name = "RBigToe"
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
id = 21
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children . children . children ] ]
name = "RSmallToe"
2024-03-06 21:11:56 +08:00
id = 23
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children . children ] ]
name = "RHeel"
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
id = 25
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children ] ]
name = "LHip"
2024-03-06 21:11:56 +08:00
id = 11
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children ] ]
name = "LKnee"
2024-03-06 21:11:56 +08:00
id = 13
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children ] ]
name = "LAnkle"
2024-03-06 21:11:56 +08:00
id = 15
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children . children ] ]
name = "LBigToe"
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
id = 20
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children . children . children ] ]
name = "LSmallToe"
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
id = 22
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children . children ] ]
name = "LHeel"
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
id = 24
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children ] ]
name = "Neck"
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
id = 18
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children ] ]
name = "Head"
Pose estimation test (#116)
Edits from @hunminkim98's awesome work at integrating pose estimation into Pose2Sim with RTMLib. Most of the changes in syntax are not necessarily better, it is mostly for the code to be more consistent with the rest of the library. Thank you again for your fantastic work!
General:
- Automatically detects whether a valid CUDA install is available. If so, use the GPU with the ONNXRuntime backend. Otherwise, use the CPU with the OpenVINO backend
- The tensorflow version used for marker augmentation was incompatible with the cuda torch installation for pose estimation: edited code and models for it to work with the latest tf version.
- Added logging information to pose estimation
- Readme.md: provided an installation procedure for CUDA (took me a while to find something simple and robust)
- Readme.md: added information about PoseEstimation with RTMLib
- added poseEstimation to tests.py
- created videos for the multi-person case (used to only have json, no video), and reorganized Demo folders. Had to recreate calibration file as well
Json files:
- the json files only saved one person, I made it save all the detected ones
- tracking was not taken into account by rtmlib, which caused issues in synchronization: fixed, waiting for merge
- took the save_to_openpose function out from the main function
- minified the json files (they take less space when all spaces are removed)
Detection results:
- Compared the triangulated locations of RTMpose keypoints to the ones of OpenPose to potentially edit model marker locations on OpenSim. Did not seem to need it.
Others in Config.toml:
- removed the "to_openpose" option, which is not needed
- added the flag: save_video = 'to_images' # 'to_video' or 'to_images' or ['to_video', 'to_images']
- changed the way frame_range was handled (made me change synchronization in depth, as well as personAssociation and triangulation)
- added the flag: time_range_around_maxspeed in synchronization
- automatically detect framerate from video, or set to 60 fps if we work from images (or give a value)
- frame_range -> time_range
- moved height and weight to project (only read for markerAugmentation, and in the future for automatic scaling)
- removed reorder_trc from triangulation and Config -> call it for markerAugmentation instead
Others:
- Provided an installation procedure for OpenSim (for the future) and made continuous installation check its install (a bit harder since it cannot be installed via pip)
- scaling from motion instead of static pose (will have to study whether it's as good or not)
- added logging to synchronization
- Struggled quite a bit with continuous integration
* Starting point of integrating RTMPose into Pose2Sim. (#111)
* RTM_to_Open
Convert format from RTMPose to OpenPose
* rtm_intergrated
* rtm_integrated
* rtm_integrated
* rtm_integrated
* rtm
* Delete build/lib/Pose2Sim directory
* rtm
* Delete build/lib/Pose2Sim directory
* Delete onnxruntime-gpu
* device = cpu
* add pose folder
* Update tests.py
* added annotation
* fix typo
* Should work be still lots of tests to run. Detailed commit coming soon
* intermediary commit
* last checks before v0.9.0
* Update continuous-integration.yml
* Update tests.py
* replaced tabs with spaces
* unittest issue
* unittest typo
* deactivated display for CI test of pose detection
* Try to make continuous integration work
* a
* b
* c
* d
* e
* f
* g
* h
* i
* j
* k
* l
---------
Co-authored-by: HunMinKim <144449115+hunminkim98@users.noreply.github.com>
2024-07-09 22:39:33 +08:00
id = 17
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children ] ]
name = "Nose"
2024-03-06 21:11:56 +08:00
id = 0
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children ] ]
name = "RShoulder"
2024-03-06 21:11:56 +08:00
id = 6
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children ] ]
name = "RElbow"
2024-03-06 21:11:56 +08:00
id = 8
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children . children ] ]
name = "RWrist"
2024-03-06 21:11:56 +08:00
id = 10
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children ] ]
name = "LShoulder"
2024-03-06 21:11:56 +08:00
id = 5
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children ] ]
name = "LElbow"
2024-03-06 21:11:56 +08:00
id = 7
2023-11-01 22:56:37 +08:00
[ [ pose . CUSTOM . children . children . children . children ] ]
name = "LWrist"
2024-03-06 21:11:56 +08:00
id = 9