WEAR: Outdoor Sports Data for Wearable and Egocentric Activity Recognition

WEAR provides synchronized wearable sensor and egocentric video data for outdoor sports activity recognition, published by Bock et al. in IMWUT in November 2024. The dataset focuses on temporal action localization and human activity recognition (HAR) in uncontrolled outdoor environments.

Use Cases

• Temporal action localization using egocentric video frames to detect activity boundaries
• Human activity recognition (HAR) using wearable sensor time-series data
• Multi-modal fusion research combining visual features with inertial measurement unit (IMU) signals

Strengths

• Multi-modal synchronization of egocentric video and wearable sensors
• Peer-reviewed publication in IMWUT Vol. 8 (2024)
• Specialized focus on outdoor sports environments rather than lab settings

Limitations

• Unknown total duration and participant count from provided metadata
• Outdoor environmental variability may introduce significant noise in video and IMU signals

Provenance

Source

Bock et al., IMWUT (Issue 4, Vol. 8, Article 175 November 2024)

Collection Method

Sensor and video recording of outdoor sports activities

Freshness

Last updated February 2026; based on research published in November 2024.