LSPe - Leeds Sports Pose Extended¶
The Leeds Sports Pose extended dataset contains 10,000 images gathered from Flickr searches for the tags ‘parkour’, ‘gymnastics’, and ‘athletics’ and consists of poses deemed to be challenging to estimate. Each image has a corresponding annotation gathered from Amazon Mechanical Turk and as such cannot be guaranteed to be highly accurate. The images have been scaled such that the annotated person is roughly 150 pixels in length. Each image has been annotated with up to 14 visible joint locations.
Use cases¶
Human body joint detection.
Properties¶
name
: leeds_sports_pose_extendedkeywords
: image_processing, detection, human_pose, keypointsdataset size
: 206,2 MBis downloadable
: yestasks
:- keypoints: (default)
primary use
: human body joint detectiondescription
: Contains image files and body parts keypoint coordinates for detecting human body joints in imagessets
: train, testmetadata file size in disk
: 473,7 kBhas annotations
: yeswhich
:- body joint keypoints
Note
This dataset is essentially the same as the leeds_sports_pose
but contains more training samples.
Metadata structure (HDF5)¶
Task: keypoints¶
/
├── train/
│ ├── image_filenames # dtype=np.uint8, shape=(11000,104) (note: string in ASCII format)
│ ├── keypoint_names # dtype=np.uint8, shape=(14,15) (note: string in ASCII format)
│ ├── keypoints # dtype=np.float, shape=(11000,14,3)
│ ├── object_fields # dtype=np.uint8, shape=(2,16) (note: string in ASCII format)
│ └── object_ids # dtype=np.int32, shape=(11000,2)
│
└── test/
├── image_filenames # dtype=np.uint8, shape=(1000,104) (note: string in ASCII format)
├── keypoint_names # dtype=np.uint8, shape=(14,15) (note: string in ASCII format)
├── keypoints # dtype=np.float, shape=(1000,14,3)
├── object_fields # dtype=np.uint8, shape=(2,16) (note: string in ASCII format)
└── object_ids # dtype=np.int32, shape=(1000,2)
Fields¶
image_filenames
: image file path+nameavailable in
: train, testdtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII format
keypoint_names
: body joint namesavailable in
: train, testdtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII format
keypoints
: keypoint coordinatesavailable in
: train, testdtype
: np.floatis padded
: Falsefill value
: -1note
: keypoint format [x1,y1,is_visible]
object_fields
: list of field names of the object id listavailable in
: train, testdtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII formatnote
: key field (field name aggregator)
object_ids
: list of field idsavailable in
: train, testdtype
: np.int32is padded
: Falsefill value
: -1note
: key field (field id aggregator)