Skip to content

Event

MotorSense releases video-EMG paired dataset

1 reports · 1 sources

Eadom Dessalene and other authors released MotorSense, which they describe as the first large-scale egocentric video and surface electromyography (EMG) paired dataset. The dataset contains synchronized video, dual-arm EMG, IMU, and audio from 100 hours of unscripted manipulation. The research uses this data for cross-modal pretraining between video and EMG, learning an action organization shared between visual and motor spaces, and transferring it to tasks such as action recognition, 3D hand reconstruction, and 3D contact reconstruction. The authors state that across four categories of action-related visual tasks and eight video datasets, this sensorimotor-aligned representation outperforms conventional video self-supervised methods.

Report timeline

  1. MotorSense releases first large-scale paired egocentric video and EMG dataset

    MotorSense is the first large-scale paired egocentric video and surface electromyography (EMG) dataset, containing synchronized video, bimanual EMG, IMU, and audio across 100 hours of unscripted manipulation. The research uses this data for cross-modal pre-training between video and EMG to learn an action organization shared across visual and motor spaces, and transfers it to tasks including action recognition, 3D hand reconstruction, and 3D contact reconstruction. Across four action-related visual tasks and eight video datasets, this sensorimotor-aligned representation outperforms conventional video self-supervised methods.

    arXiv cs.RO ·

Related topics

Grasping and manipulationRobot models and learningData and simulationResearch papers

Interest trend

Current
3
Peak
4Nov 4 01:00
Past 24 hours
–

The trend compares participants observed throughout the same period, so its range may be smaller than the current interest count. Move or tap to inspect each hour; use the arrow keys on a keyboard.