deepannotateaicc-by-4.0
Cattle Farming Multimodal Dataset v1
An egocentric agricultural dataset capturing fine-grained hand interactions, hand poses, and activity recognition during cattle care and animal interaction operations. It includes RGB-D video, 2D/3D hand keypoints, hand-object interactions, object annotations, motion statistics, synthetic IMU data, and natural language semantic annotations.
Downloads1
Episodes3 activity sequences (300 frames total across Animal_Farming_1, Animal_Farming_2, and Cattle_Farming)
Why This Matters for Physical AI
This dataset advances embodied AI research in agricultural and livestock management by providing fine-grained egocentric hand-object interaction annotations with multimodal sensor data, enabling training of action-recognition and manipulation systems for farm automation.
Technical Profile
- Modalities
- rgbdepthhand_pose_2dhand_pose_3dimulanguagebounding_boxesmotion_statistics
- Action Space
- language
- Environment
- farmagriculture
- Task Types
- human_object_interactionaction_recognitionhand_pose_estimationanimal_care
- Episodes
- 3 activity sequences (300 frames total across Animal_Farming_1, Animal_Farming_2, and Cattle_Farming)
- Data Format
- HDF5
- Annotation Types
- hand_keypoints_2dhand_keypoints_3dhand_object_interactionsobject_bounding_boxesaction_labelslanguage_instructionsmotion_statisticstrajectory_annotations
- License
- cc-by-4.0
Access
Need custom rgb data?
Claru builds purpose-built datasets for farm applications with dense human annotations and quality assurance.
Request a Sample Pack