Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
4
datasets available to search
ShareScore release 0.9.0
Dataset results
4 results for “Hand gesture recognition”
EMG and Video Dataset for sensor fusion based hand gestures recognition
<p>This dataset contains data for hand gesture recognition recorded with 3 different sensors. </p> <p>sEMG: recorded via the Myo armband that is composed of 8 equally spaced non-invasive sEMG sensors that can be placed approximately around the middle of the forearm. The sampling frequency of Myo is 200 Hz. The output of the Myo is a.u </p> <p>DVS: Dynamic Video Sensor which is a very low power event-based camera with 128x128 resolution</p> <p>DAVIS: Dynamic Video Sensor which is a very low power event-based camera with 240x180 resolution that also acquires APS frames.</p> <p>The dataset contains recordings of 21 subjects. Each subject performed 3 sessions, where each of the 5 hand gesture was recorded 5 times, each lasting for 2s. Between the gestures a relaxing phase of 1s is present where the muscles could go to the rest position, removing any residual muscular activation.</p> <p> </p> <p>Note: All the information for the DVS sensor has been extracted and can be found in the *.npy files. In case the raw data (.aedat) was needed please contact</p> <p> </p> <p>enea.ceolini@ini.uzh.ch</p> <p>elisa@ini.uzh.ch</p> <p>==== README ====</p> <p> </p> <p>DATASET STRUCTURE:</p> <p>EMG, DVS and APS recordings</p> <p>21 subjects</p> <p>3 sessions for each subject</p> <p>5 gestures in each session ('pinky', 'elle', 'yo', 'index', 'thumb')</p> <p> </p> <p>SINGLE DATASETS:</p> <p>- relax21_raw_emg.zip: contains raw sEMG and annotations (ground truth of gestures) in the format `subjectXX_sessionYY_ZZZ` with `XX` subject ID (01 to 21), `YY` session ID (01-03) and `ZZZ` that can be ‘emg’ or ‘ann’.</p> <p> </p> <p>- relax21_raw_dvs.zip: contains the full-frame dvs events in an array with dimensions 0 -> addr_x, 1 -> addr_y, 2 -> timestamp, 3 -> polarity. The timestamps are in seconds and synchronized with the Myo. Each file is in the format `subjectXX_sessionYY_dvs` with `XX` subject ID (01 to 21), `YY` session ID (01-03).</p> <p> </p> <p>- relax21_cropped_aps.zip: contains the 40x40 pixel aps frames for all subjects and trials in the format `subjectXX_sessionYY_Z_W_K` with `XX` subject ID (01 to 21), `YY` session ID (01-03), Z gesture ('pinky', 'elle', 'yo', 'index', 'thumb’), W trial ID (1-5), `K` frame index.</p> <p> </p> <p>- relax21_cropped_dvs_emg_spikes.pkl: spiking dataset that can be used to reproduce the results in the paper. The dataset is a dictionary with the following keys:</p> <ul> <li><strong>- </strong><strong>y</strong>: array of size 1xN with the class (0->4).</li> <li><strong>- </strong><strong>sub</strong>: array of size 1xN with the subject id (1->10).</li> <li><strong>- </strong><strong>sess</strong>: array of size 1xN with the session id (1->3).</li> <li><strong>- </strong><strong>dvs</strong>: list of length N, each object in the list is a 2d array of size 4xT_n where T_n is the number of events in the trial and the 4 dimensions rappresent: 0 -> addr_x, 1 -> addr_y, 2 -> timestamp, 3 -> polarity .</li> <li><strong>- </strong><strong>emg</strong>: list of length N, each object in the list is a 2d array of size 3xT_n where T_n is the number of events in the trial and the 3 dimensions rappresent: 0 -> addr, 1 -> timestamp, 3 -> polarity.</li> </ul> <p> </p> <p> </p>
RIS Based Hand Gesture Recognition Dataset
<h1><strong>RIS Based Hand Gesture Recognition Dataset</strong></h1> <h2><strong>Overview</strong></h2> <div>This dataset contains images for gesture recognition, divided into two main sets: dataset0608 and data_synthetic_variab. The data was collected using a wooden hand. </div> <div> </div> <h3>dataset0608</h3> <div>This dataset consists of two modes: ris_random and ris_optimized. The main difference between the two subfolders is the configuration of the RIS (random or optimized).</div> <div> </div> <div>This dataset consists of four subfolders: ris_random, ris_random2, ris_optimized, and ris_optimized2. The main difference between the subfolders is the format of the data:</div> <div>- ris_random and ris_optimized: Data is stored in individual files for each frame, named as 'frame_{i}{posture}{n_med}' </div> <div>- ris_random2 and ris_optimized2: Data has already been processed and combined into single files for all frames using the compact_files_frames.txt function, named as 'all_frames_{posture}_{n_med}' </div> <div> </div> <div>For each gestures = {close, two, open}, we have n_med values from 0 to 114 and 10 frames. Therefore, the ris_random and ris_optimized folders contain 10 frames × 115 measurements × 3 gestures = 3450 files, while the ris_random2 and ris_optimized2 folders contain 1 × 115 measurements × 3 gestures = 345 files.</div> <div> </div> <h3>data_synthetic_variab</h3> <div>This dataset consists of two modes: ris_random and ris_optimized. The main difference between the two subfolders is the configuration of the RIS (random or optimized). </div> <div> </div> <div>This dataset consists of four subfolders: ris_random, ris_random2, ris_optimized, and ris_optimized2. The main difference between the subfolders is the format of the data:</div> <div>- ris_random and ris_optimized: Data is stored in individual files for each frame, named as 'frame_{i}{posture}{n_med}' </div> <div>- ris_random2 and ris_optimized2: Data has already been processed and combined into single files for all frames using the compact_files_frames.txt function, named as 'all_frames_{posture}_{n_med}' </div> <div> </div> <div>For each gestures = {close, two, open}, we have n_med values from 0 to 8 and 10 frames. This dataset provides additional synthetic data with variations in hand position to increase the dataset's diversity. Each gesture is represented by 8 different ways, where the hand position was slightly modified between each sample. These real data were used as a basis for generating synthetic data. By using the functions in the files "multiply_files.txt" and "add_gaussian_noise.txt," the dataset was expanded and made more realistic by adding Gaussian noise to the images.</div> <div> </div> <div>Therefore, the ris_random and ris_optimized folders contain 10 frames × 8 measurements × 3 gestures = 240 files, while the ris_random2 and ris_optimized2 folders contain 1 × 8 measurements × 3 gestures = 24 files.</div> <div> </div> <h3>Functions</h3> <div>* **add_gaussian_noise.txt:** This script adds Gaussian noise to the images to simulate real-world conditions and improve the robustness of the model.</div> <div>* **compact_files_frames.txt:** This script combines multiple frames into a single image, which can be useful for certain types of analysis.</div>
EMG from forearm datasets for hand gestures recognition
<p>This dataset contains 2 sets of sEMG recordings: a set containing <strong>PINCH</strong> movements (4 pinches between thumb and index/middle/ring/pinky finger) and a set containing <strong>ROSHAMBO</strong> movements (3 movements: rock, paper, scissors). Both sets have been recorded with the Myo armband. The Myo is composed of 8 equally spaced non-invasive sEMG sensors that can be placed approximately around the middle of the forearm. The sampling frequency of Myo is 200 Hz. The output of the Myo is a.u.. The <strong>PINCH </strong>set contains recordings of 22 subjects whilst the <strong>ROSHAMBO</strong> set contains recordings of 10 subjects. Each subject performed 3 sessions, where each hand gesture was recorded 5 times, each lasting for 2s. Between the gestures a relaxing phase of 1s is present where the muscles could go to the rest position, removing any residual muscular activation. Full details for the <strong>ROSHAMBO</strong> set can be found in:</p> <p>Donati, Elisa, et al. "Processing EMG signals using reservoir computing on an event-based neuromorphic system." <em>2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)</em>. IEEE, 2018.</p> <p>For each session, the dataset contains 2 <strong>*.npy</strong> files one specifying the EMG data (<strong>*_emg.npy</strong>) the other one (<strong>*_ann.npy</strong>) specifying the corresponding gestures along the sampled EMG. The data can be easily loaded in python with numpy.</p> <p> </p>
Dataset for Dynamic Hand Gesture Recognition Systems
<p>Computer vision systems are commonly used to design touchless human-computer interfaces (HCI) based on dynamic hand gesture recognition (HGR) systems, which have a wide range of applications in several domains, such as gaming, multimedia, automotive, and home automation. However, automatic HGR is still a challenging task, mostly because of the diversity in how people perform the gestures. In addition, the number of publicly available hand gesture datasets is scarce; often, the gestures are not acquired with sufficient image quality, and the gestures are not correctly performed. In this data article, we propose a dataset of 27 dynamic hand gesture types acquired at full HD resolution from 21 different subjects, which were carefully instructed before performing the gestures and monitored when performing the gesture; the subjects had to repeat the movement in case the performed hand gesture was not correct, i.e., the authors of this paper that were observing the gesture found that it did not correspond to the exact expected movement and/or the camera recorded a viewpoint did not allow for a plain visualizing of the gesture. Each subject performed 3 times the 27 hand gestures for a total of 1701 videos collected and corresponding to 204120 video frames.</p> <p> </p> <p>In the following, we discuss the details of the provided datasets.</p> <p><strong>hand_gestures_dataset_videos.zip</strong> - This dataset contains the videos of the recorded hand gestures. The zip contains 27 main folders. Each main folder refers to a hand gesture class, for a total of 27 main folders named “class_xx”, where “xx” identifies the class from 01 to 27. Within each of the class folders, there are 21 sub-folders, one folder for each of the subjects that performed the hand gestures. These folders are named “Useryy_”, where “yy” identifies the user from 01 to 21. Each of the user folders contains three videos (.avi) corresponding to the three hand gestures performed by the user for each hand gesture class. The size of the full dataset is 21.34 GB.</p> <p><strong>HGD_VideoFrames_class_XX.zip</strong> - These datasets contain the video frames extracted from the videos of the recorded hand gestures. Each zip file contains the video frames of a hand gesture class, for a total of 27 zip files named “HGD_VideoFrames_class_XX.zip”, where “xx” identifies the class from 01 to 27. Therefore, each zip file contains one of the 27 class folders. Within each of the class folders, there are 21 sub-folders, one folder for each of the subjects that performed the hand gestures. These folders are named “Useryy_”, where “yy” identifies the user from 01 to 21. Each of the user folders contains, in turn, 3 sub-folders, one folder for each of the three hand gestures performed for each hand gesture class. These sub-folders are named, respectively, “Useryy_1”, “Useryy_2”, and “Useryy_3”, and contain 120 video frames (.png) extracted from the corresponding video. The size of each zip file is about 9 GB.</p> <p><strong>hand_gesture_timing_stats.csv</strong> - This dataset contains timing information regarding the gestures performed by the subjects. The size of this dataset is 36 KB. It has 567 records plus the header. The meaning of the columns is as follows:</p> <ul> <li> <p><em>class</em>: hand gesture class, from 01 to 27.</p> </li> <li> <p><em>user</em>: user who performed the hand gestures, from 01 to 21.</p> </li> <li> <p><em>start_frame_1</em>: starting frame related to the first performed hand gesture. It is a number between 0 and 119.</p> </li> <li> <p><em>end_frame_1</em>: ending frame related to the first performed hand gesture. It is a number between 0 and 119.</p> </li> <li> <p><em>exec_time_1</em>: execution time (in seconds) related to the first performed hand gesture. It is computed as the difference between the ending frame and the starting frame divided by the 30 fps set for video recording.</p> </li> <li> <p><em>start_frame_2</em>: starting frame related to the second performed hand gesture. It is a number between 0 and 119.</p> </li> <li> <p><em>end_frame_2</em>: ending frame related to the second performed hand gesture. It is a number between 0 and 119.</p> </li> <li> <p><em>exec_time_2</em>: execution time (in seconds) related to the second performed hand gesture. It is computed as the difference between the ending frame and the starting frame divided by the 30 fps set for video recording.</p> </li> <li> <p><em>start_frame_3</em>: starting frame related to the third performed hand gesture. It is a number between 0 and 119.</p> </li> <li> <p><em>end_frame_3</em>: ending frame related to the third performed hand gesture. It is a number between 0 and 119.</p> </li> <li> <p><em>exec_time_3</em>: execution time (in seconds) related to the third performed hand gesture. It is computed as the difference between the ending frame and the starting frame divided by the 30 fps set for video recording.</p> </li> <li> <p><em>mean_exec_time</em>: mean execution time (in seconds) for that related user and hand gesture. It is computed as the mean of the execution times computed for the three hand gestures performed by that user for that class.</p> </li> <li> <p><em>std_dev_exec_time</em>: standard deviation of the three execution times (in seconds) computed for the three hand gestures performed by that user for that class.</p> </li> <li> <p><em>total_mean_exec_time</em>: total mean execution time (in seconds) for that class. It is computed as the mean of all the execution times computed for the three hand gestures performed by all the users for that class.</p> </li> <li> <p><em>total_std_dev_exec_time</em>: total standard deviation of all the execution times (in seconds) for that class. It is computed as the standard deviation of all the execution times computed for the three hand gestures performed by all the users for that class. Note that in this case, the standard deviation has been computed by dividing by (N-1) as the entire population is considered.</p> </li> </ul> <p> </p> <p dir="ltr"><strong>If you make use of this dataset, please consider citing the following publication:</strong></p> <p dir="ltr">Fronteddu, G., Porcu, S., Floris, A., & Atzori, L. (2022). A dynamic hand gesture recognition dataset for human-computer interfaces. Computer Networks, 205, 108781.</p> <p dir="ltr">BibTex format:</p> <p>@article{fronteddu2022dynamic, title={A dynamic hand gesture recognition dataset for human-computer interfaces}, author={Fronteddu, Graziano and Porcu, Simone and Floris, Alessandro and Atzori, Luigi}, journal={Computer Networks}, volume={205}, pages={108781}, year={2022}, publisher={Elsevier}, doi = {https://doi.org/10.1016/j.comnet.2022.108781} }</p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.