Gym Workouts Recognition Dataset with IMU and Capacitive Sensors

RecGym contains sensor data for recognizing gym exercises, collected from inertial measurement units (IMU) and capacitive sensors. The dataset supports multimodal and time-series analysis for workout recognition tasks. It originates from the UCI Machine Learning Repository.

Use Cases

• Classify workout types using fused IMU and capacitive sensor time-series data.
• Train sequence models like LSTMs on the multimodal sensor streams for real-time activity recognition.
• Benchmark sensor fusion techniques by comparing performance using IMU features versus capacitive sensor features.
• Develop personal fitness trackers that can identify specific gym exercises from the sensor patterns.

Strengths

• Includes data from two distinct sensor modalities (IMU and capacitive).
• Specifically designed for gym workout recognition, a focused domain.

Limitations

• Unknown sample size and number of participants limits assessment of statistical power.
• Lack of detailed column information prevents evaluation of feature completeness.

Provenance

Source

UCI Machine Learning Repository

Collection Method

Collected using wearable IMU and capacitive sensors during gym workouts.