Welcome to Ninaweb


We are currently restructuring the website.
There might be temporary access problems to some content.



In Ninaweb you can find:

1. The Ninapro database

Ninapro is a publicly available multimodal database to foster research on human, robotic & prosthetic hands and on machine learning based control systems.
The database includes over 300 data acquisitions divided in 10 datasets, providing electromyography, kinematic, inertial, eye tracking, visual, clinical and neurocognitive data and targeting different research needs.
Ninapro datasets are used worldwide by scientific researchers in machine learning, robotics, medical and neurocognitive sciences.
Intact subjects can be used as a proxy measure for hand amputees (Atzori et al., EMBC 2014), however we recommend to include amputees data whenever possible.
Considering the data from amputees, several clinical parameters related to the amputation can significantly influence results, as described in Atzori et al., Journal of Rehabilitation Research and Development, 2016. This fact should be properly considered while analyzing the data and presenting the results.

Ninapro database in brief:
DB1 - Subjects: 27; Data: EMG (Delsys Trigno), kinematics.
DB2 - Subjects: 40; Data: EMG (Delsys Trigno), kinematics, inertial, force.
DB3 - Subjects: 11 transradial amputees; Data: EMG, inertial.
DB4 - Subjects: 10; EMG (Cometa electrodes), kinematics.
DB5 - Subjects: 10; Data: EMG, inertial (two talmic Myo Armbands, 16 electrodes)
DB6 - Repeatability data (2 times per day acquisitions for 5 days from 10 subjects); data: EMG, inertial, portable eye tracking glasses with scene camera
DB7 - Subjects: 20 intact subjects, 2 transradial amputees; Data: EMG
DB8 - aimed at the estimation/reconstruction of finger movement rather than motion/grip classification. Subjects: 10 intact subjects and 2 amptuees.
DB9 - Subjects: 77; Data: kinematic data acquired with a Cyberglove-II.
DB10, MeganePro - Subjects: 30 amputee and 30 intact subjects; Data: sEMG, intertial, gaze tracking, visual, behavioral and clinical data for prosthetics and the analysis of phantom limb sensation.

2. Resources, results and projects information

Our extremely fast signal feature extraction tool, allowing to extract the most common signal features in minutes from dozens of subjects (currently implemented for Matlab).

Quantitative taxonomy of hand grasps
The most solid representation of hand movements' hierarchy, fully based on quantitative EMG and kinematics data analysis (image below).
The taxonomy is useful for rehabilitation, robotics, virtual reality and other domains and it is published on JNER (Stival et al, JNER 2019) .

Synergies & their variability across subjects or acquisitions
Variability of muscle synergies in hand grasps: analysis of intra-and inter-session data, Sensors, 2019
Kinematic synergies of hand grasps on 77 subjects, JNER, 2019
Muscle synergies of hand grasps on 28 subjects - Frontiers in Neurorobotics, 2018

Variability of sEMG analysis across subjects or acquisitions
Variability of muscle synergies in hand grasps: analysis of intra-and inter-session data, Sensors, 2019
Questioning domain adaptation in myoelectric hand prostheses control, Sensors, 2021
Repeatability of grasp recognition for robotic hand prosthesis control based on sEMG data, ICORR, 2017

An innovative robotic hand prosthesis controlled with machine learning and produced with additive manufacturing.
Click here to watch a video about our advancements on the Swiss National Italian television.

We hope that you will enjoy Ninaweb.
Please, contact us if you would like to get involved into one of our projects.

The Ninaweb team