UCF Sports Action¶
UCF Sports dataset consists of a set of actions collected from various sports which are typically featured on broadcast television channels such as the BBC and ESPN. The video sequences were obtained from a wide range of stock footage websites including BBC Motion gallery and GettyImages.
The dataset includes a total of 150 sequences with the resolution of 720 x 480. The collection represents a natural pool of actions featured in a wide range of scenes and viewpoints.
Use cases¶
Human action recognition in videos.
Properties¶
name
: ucf_sportskeywords
: image_processing, recognition, detection, activity, human, single_persondataset size
: 1,8 GBis downloadable
: yestasks
:- recognition: (default)
primary use
: action recognition in videosdescription
: Contains videos and action label annotations for action recognitionsets
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05metadata file size in disk
: 1,0 MBhas annotations
: yeswhich
:- activity labels for each video.
Metadata structure (HDF5)¶
Task: recognition¶
/
├── train01/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(6548,4)
│ ├── image_filenames # dtype=np.uint8, shape=(6548,74) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(103,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(6548,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(103,127)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(103,127)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(103,127)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,15)
│
├── test01/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(3032,4)
│ ├── image_filenames # dtype=np.uint8, shape=(3032,76) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(47,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(3032,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(47,144)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(47,144)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(47,144)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,7)
│
├── train02/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(6529,4)
│ ├── image_filenames # dtype=np.uint8, shape=(6529,76) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(103,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(6529,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(103,144)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(103,144)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(103,144)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,15)
│
├── test02/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(3051,4)
│ ├── image_filenames # dtype=np.uint8, shape=(3051,74) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(47,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(3051,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(47,123)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(47,123)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(47,123)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,7)
│
├── train03/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(6537,4)
│ ├── image_filenames # dtype=np.uint8, shape=(6537,74) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(103,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(6537,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(103,144)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(103,144)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(103,144)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,15)
│
├── test03/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(3034,4)
│ ├── image_filenames # dtype=np.uint8, shape=(3034,76) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(47,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(3034,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(47,127)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(47,127)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(47,127)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,7)
│
├── train04/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(6520,4)
│ ├── image_filenames # dtype=np.uint8, shape=(6520,74) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(103,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(6520,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(103,127)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(103,127)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(103,127)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,15)
│
├── test04/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(3060,4)
│ ├── image_filenames # dtype=np.uint8, shape=(3060,73) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(47,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(3060,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(47,144)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(47,144)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(47,144)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,7)
│
├── train05/
│ ├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
│ ├── boxes # dtype=np.int32, shape=(6542,4)
│ ├── image_filenames # dtype=np.uint8, shape=(6542,76) (note: string in ASCII format)
│ ├── videos # dtype=np.uint8, shape=(103,24) (note: string in ASCII format)
│ ├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(6542,4)
│ ├── list_boxes_per_video # dtype=np.int32, shape=(103,144)
│ ├── list_filenames_per_video # dtype=np.int32, shape=(103,144)
│ ├── list_object_ids_per_video # dtype=np.int32, shape=(103,144)
│ └── list_videos_per_activity # dtype=np.int32, shape=(10,15)
│
└── test05/
├── activities # dtype=np.uint8, shape=(10,14) (note: string in ASCII format)
├── boxes # dtype=np.int32, shape=(3038,4)
├── image_filenames # dtype=np.uint8, shape=(3038,75) (note: string in ASCII format)
├── videos # dtype=np.uint8, shape=(47,24) (note: string in ASCII format)
├── object_fields # dtype=np.uint8, shape=(4,16) (note: string in ASCII format)
├── object_ids # dtype=np.int32, shape=(3038,4)
├── list_boxes_per_video # dtype=np.int32, shape=(47,127)
├── list_filenames_per_video # dtype=np.int32, shape=(47,127)
├── list_object_ids_per_video # dtype=np.int32, shape=(47,127)
└── list_videos_per_activity # dtype=np.int32, shape=(10,7)
Fields¶
activities
: activity namesavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII format
image_filenames
: image file path+nameavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII format
boxes
: bounding box coordinatesavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.int32is padded
: Falsefill value
: -1note
: bbox format [x1,y1,x2,y2]
videos
: video nameavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII format
object_fields
: list of field names of the object id listavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII formatnote
: key field (field name aggregator)
object_ids
: list of field idsavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.int32is padded
: Falsefill value
: -1note
: key field (field id aggregator)
list_boxes_per_video
: list of bounding box ids per videoavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.int32is padded
: Truefill value
: -1note
: pre-ordered list
list_filenames_per_video
: list of image ids per videoavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.int32is padded
: Truefill value
: -1note
: pre-ordered list
list_object_ids_per_video
: list of object ids per videoavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.int32is padded
: Truefill value
: -1note
: pre-ordered list
list_videos_per_activity
: list of video ids per activityavailable in
: train01, train02, train03, train04, train05, test01, test02, test03, test04, test05dtype
: np.int32is padded
: Truefill value
: -1note
: pre-ordered list