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.
18
datasets available to search
ShareScore release 0.9.0
Dataset results
18 results for “Human Activity Recognition”
Wallhack1.8k Dataset | Data Augmentation Techniques for Cross-Domain WiFi CSI-Based Human Activity Recognition
<p>This repository contains the <strong>Wallhack1.8k dataset</strong> for WiFi-based long-range activity recognition in Line-of-Sight (LoS) and Non-Line-of-Sight (NLoS)/Through-Wall scenarios, as proposed in [1,2], as well as the <strong>CAD models</strong> (of 3D-printable parts) of the WiFi systems proposed in [2].</p> <p><strong>PyTroch Dataloader</strong></p> <p>A minimal PyTorch dataloader for the Wallhack1.8k dataset is provided at: <a href="https://github.com/StrohmayerJ/wallhack1.8k" target="_blank" rel="noopener">https://github.com/StrohmayerJ/wallhack1.8k</a></p> <p><strong>Dataset Description</strong></p> <p>The Wallhack1.8k dataset comprises 1,806 CSI amplitude spectrograms (and raw WiFi packet time series) corresponding to three activity classes: "no presence," "walking," and "walking + arm-waving." WiFi packets were transmitted at a frequency of 100 Hz, and each spectrogram captures a temporal context of approximately 4 seconds (400 WiFi packets).</p> <p>To assess cross-scenario and cross-system generalization, WiFi packet sequences were collected in LoS and through-wall (NLoS) scenarios, utilizing two different WiFi systems (BQ: biquad antenna and PIFA: printed inverted-F antenna). The dataset is structured accordingly:</p> <ul> <li>LOS/BQ/ <- WiFi packets collected in the LoS scenario using the BQ system</li> <li>LOS/PIFA/ <- WiFi packets collected in the LoS scenario using the PIFA system</li> <li>NLOS/BQ/ <- WiFi packets collected in the NLoS scenario using the BQ system</li> <li>NLOS/PIFA/ <- WiFi packets collected in the NLoS scenario using the PIFA system</li> </ul> <p>These directories contain the raw WiFi packet time series (see Table 1). Each row represents a single WiFi packet with the complex CSI vector <em>H</em> being stored in the "data" field and the class label being stored in the "class" field. <em>H </em>is of the form [I, R, I, R, ..., I, R], where two consecutive entries represent imaginary and real parts of complex numbers (the Channel Frequency Responses of subcarriers). Taking the absolute value of <em>H</em> (e.g., via <em>numpy.abs(H)</em>) yields the subcarrier amplitudes <em>A</em>.</p> <p>To extract the 52 L-LTF subcarriers used in [1], the following indices of <em>A</em> are to be selected:</p> <pre><code># 52 L-LTF subcarriers csi_valid_subcarrier_index = [] csi_valid_subcarrier_index += [i for i in range(6, 32)] csi_valid_subcarrier_index += [i for i in range(33, 59)]</code></pre> <p>Additional 56 HT-LTF subcarriers can be selected via:</p> <pre><code># 56 HT-LTF subcarriers csi_valid_subcarrier_index += [i for i in range(66, 94)] csi_valid_subcarrier_index += [i for i in range(95, 123)]</code></pre> <p>For more details on subcarrier selection, see <a href="https://docs.espressif.com/projects/esp-idf/en/stable/esp32/api-guides/wifi.html">ESP-IDF</a> (Section Wi-Fi Channel State Information) and <a href="https://github.com/espressif/esp-csi">esp-csi</a>.</p> <p>Extracted amplitude spectrograms with the corresponding label files of the train/validation/test split: "trainLabels.csv," "validationLabels.csv," and "testLabels.csv," can be found in the <em>spectrograms/</em> directory.</p> <p>The columns in the label files correspond to the following: [Spectrogram index, Class label, Room label]</p> <ul> <li>Spectrogram index: [0, ..., n]</li> <li>Class label: [0,1,2], where 0 = "no presence", 1 = "walking", and 2 = "walking + arm-waving."</li> <li>Room label: [0,1,2,3,4,5], where labels 1-5 correspond to the room number in the NLoS scenario (see Fig. 3 in [1]). The label 0 corresponds to no room and is used for the "no presence" class.</li> </ul> <p><strong>Dataset Overview:</strong></p> <p>Table 1: Raw WiFi packet sequences.</p> <table> <tbody> <tr> <td><strong>Scenario</strong></td> <td><strong>System</strong></td> <td><em>"no presence" / label 0</em></td> <td><em>"walking" / label 1</em></td> <td><em>"walking + arm-waving" / label 2</em></td> <td><strong>Total</strong></td> </tr> <tr> <td>LoS</td> <td>BQ</td> <td>b1.csv</td> <td>w1.csv, w2.csv, w3.csv, w4.csv and w5.csv</td> <td>ww1.csv, ww2.csv, ww3.csv, ww4.csv and ww5.csv</td> <td> </td> </tr> <tr> <td>LoS</td> <td>PIFA</td> <td>b1.csv</td> <td>w1.csv, w2.csv, w3.csv, w4.csv and w5.csv</td> <td>ww1.csv, ww2.csv, ww3.csv, ww4.csv and ww5.csv</td> <td> </td> </tr> <tr> <td>NLoS</td> <td>BQ</td> <td>b1.csv</td> <td>w1.csv, w2.csv, w3.csv, w4.csv and w5.csv</td> <td>ww1.csv, ww2.csv, ww3.csv, ww4.csv and ww5.csv</td> <td> </td> </tr> <tr> <td>NLoS</td> <td>PIFA</td> <td>b1.csv</td> <td>w1.csv, w2.csv, w3.csv, w4.csv and w5.csv</td> <td>ww1.csv, ww2.csv, ww3.csv, ww4.csv and ww5.csv</td> <td> </td> </tr> <tr> <td> </td> <td> </td> <td>4</td> <td>20</td> <td>20</td> <td><strong>44</strong></td> </tr> </tbody> </table> <p>Table 2: Sample/Spectrogram distribution across activity classes in Wallhack1.8k.</p> <table> <tbody> <tr> <td><strong>Scenario</strong></td> <td><strong>System</strong></td> <td> <p><em>"no presence" / </em> label 0</p> </td> <td> <p><em>"walking"</em> / label 1</p> </td> <td><em>"walking + arm-waving" / </em>label 2</td> <td><strong>Total</strong></td> </tr> <tr> <td>LoS</td> <td>BQ</td> <td>149</td> <td>154</td> <td>155</td> <td> </td> </tr> <tr> <td>LoS</td> <td>PIFA</td> <td>149</td> <td>160</td> <td>152</td> <td> </td> </tr> <tr> <td>NLoS</td> <td>BQ</td> <td>148</td> <td>150</td> <td>152</td> <td> </td> </tr> <tr> <td>NLoS</td> <td>PIFA</td> <td>143</td> <td>147</td> <td>147</td> <td> </td> </tr> <tr> <td> </td> <td> </td> <td>589</td> <td>611</td> <td>606</td> <td><strong>1,806</strong></td> </tr> </tbody> </table> <p> </p> <p><strong>Download and Use</strong><br>This data may be used for non-commercial research purposes only. If you publish material based on this data, we request that you include a reference to one of our papers [1,2].</p> <p>[1] Strohmayer, Julian, and Martin Kampel. (2024). “Data Augmentation Techniques for Cross-Domain WiFi CSI-Based Human Activity Recognition”, <em>In IFIP International Conference on Artificial Intelligence Applications and Innovations</em> (pp. 42-56). Cham: Springer Nature Switzerland<em>,</em> doi: <a href="https://doi.org/10.1007/978-3-031-63211-2_4" target="_blank" rel="noopener">https://doi.org/10.1007/978-3-031-63211-2_4</a>.</p> <p>[2] Strohmayer, Julian, and Martin Kampel., “Directional Antenna Systems for Long-Range Through-Wall Human Activity Recognition,” <em>2024 IEEE International Conference on Image Processing (ICIP)</em>, Abu Dhabi, United Arab Emirates, 2024, pp. 3594-3599, doi: <a href="https://doi.org/10.1109/ICIP51287.2024.10647666" target="_blank" rel="noopener">https://doi.org/10.1109/ICIP51287.2024.10647666</a>.</p> <p>BibTeX citations:</p> <pre>@inproceedings{strohmayer2024data, title={Data Augmentation Techniques for Cross-Domain WiFi CSI-Based Human Activity Recognition}, author={Strohmayer, Julian and Kampel, Martin}, booktitle={IFIP International Conference on Artificial Intelligence Applications and Innovations}, pages={42--56}, year={2024}, organization={Springer}}<br><br>@INPROCEEDINGS{10647666,<br> author={Strohmayer, Julian and Kampel, Martin},<br> booktitle={2024 IEEE International Conference on Image Processing (ICIP)}, <br> title={Directional Antenna Systems for Long-Range Through-Wall Human Activity Recognition}, <br> year={2024},<br> volume={},<br> number={},<br> pages={3594-3599},<br> keywords={Visualization;Accuracy;System performance;Directional antennas;Directive antennas;Reflector antennas;Sensors;Human Activity Recognition;WiFi;Channel State Information;Through-Wall Sensing;ESP32},<br> doi={10.1109/ICIP51287.2024.10647666}}<br><br><br></pre>
Dataset: WiFi-based Human Activity Recognition using Raspberry Pi
<p>This dataset contains 980 802.11 Channel State Information captures for 11 activities performed in a small apartment by 1 subject. For full description, check README.md.</p>
WiFi CSI-Based Long-Range Through-Wall Human Activity Recognition with the ESP32
<p><strong>WiFi CSI-Based Long-Range Through-Wall Human Activity Recognition with the ESP32</strong></p> <p>This repository contains the WiFi CSI human presence detection and activity recognition datasets proposed in [1].</p> <p><strong>Datasets</strong></p> <ul> <li><strong>DP_LOS</strong> - Line-of-sight (LOS) presence detection dataset, comprised of 392 CSI amplitude spectrograms.</li> <li><strong>DP_NLOS </strong>- Non-line-of-sight (NLOS) presence detection dataset, comprised of 384 CSI amplitude spectrograms.</li> <li><strong>DA_LOS</strong> - LOS activity recognition dataset, comprised of 392 CSI amplitude spectrograms.</li> <li><strong>DA_NLOS</strong> - NLOS activity recognition dataset, comprised of 384 CSI amplitude spectrograms.</li> </ul> <p>Table 1: Characteristics of presence detection and activity recognition datasets. </p> <table> <tbody> <tr> <td><strong>Dataset</strong></td> <td><strong>Scenario</strong></td> <td><strong>#Rooms</strong></td> <td><strong>#Persons</strong></td> <td><strong>#Classes</strong></td> <td><strong>Packet Sending Rate</strong></td> <td><strong>Interval </strong></td> <td><strong>#Spectrograms</strong></td> </tr> <tr> <td>DP_LOS</td> <td>LOS</td> <td>1</td> <td>1</td> <td>6</td> <td>100Hz</td> <td>4s (400 packets)</td> <td>392</td> </tr> <tr> <td>DP_NLOS</td> <td>NLOS</td> <td>5</td> <td>1</td> <td>6</td> <td>100Hz</td> <td>4s (400 packets)</td> <td>384</td> </tr> <tr> <td>DA_LOS</td> <td>LOS</td> <td>1</td> <td>1</td> <td>3</td> <td>100Hz</td> <td>4s (400 packets)</td> <td>392</td> </tr> <tr> <td>DA_NLOS</td> <td>NLOS</td> <td>5</td> <td>1</td> <td>3</td> <td>100Hz</td> <td>4s (400 packets)</td> <td>384</td> </tr> </tbody> </table> <p> </p> <p><strong>Data Format</strong></p> <p>Each dataset employs an 8:1:1 training-validation-test split, defined in the provided label files <em>trainLabels.csv</em>, <em>validationLabels.csv</em>, and <em>testLabels.csv</em>. Label files use the sample format [<em>i c</em>], with <em>i</em> corresponding to the spectrogram index (i.png) and <em>c </em>corresponding to the class. For presence detection datasets (DP_LOS <em>, </em>DP_NLOS), c in {0 = "no presence", 1 = "presence in room 1", ..., 5 = "presence in room 5"}. For activity recognition datasets (DA_LOS <em>, </em>DA_NLOS), c in {0="no activity", 1="walking", and 2="walking + arm-waving"}. Furthermore, the mean and standard deviation of a given dataset are provided in <em>meanStd.csv</em>.</p> <p><strong>Download and Use</strong><br>This data may be used for non-commercial research purposes only. If you publish material based on this data, we request that you include a reference to our paper [1].</p> <p>[1] Strohmayer, Julian, and Martin Kampel. "WiFi CSI-Based Long-Range Through-Wall Human Activity Recognition with the ESP32" <em>International Conference on Computer Vision Systems</em>. Cham: Springer Nature Switzerland, 2023. </p> <p>BibTeX citation:</p> <pre>@inproceedings{strohmayer2023wifi, title={WiFi CSI-Based Long-Range Through-Wall Human Activity Recognition with the ESP32}, author={Strohmayer, Julian and Kampel, Martin}, booktitle={International Conference on Computer Vision Systems}, pages={41--50}, year={2023}, organization={Springer} }</pre>
The FORTH-TRACE dataset for human activity recognition of simple activities and postural transitions using a Body Area Network
<p>The dataset is collected from 15 participants wearing 5 Shimmer wearable sensor nodes on the locations listed in Table 1. The participants performed a series of 16 activities (7 basic and 9 postural transitions), listed in Table 2.</p> <p>The captured signals are the following:</p> <ul> <li>3-axis accelerometer</li> <li>3-axis gyroscope</li> <li>3-axis magnetometer</li> </ul> <p>The sampling rate of the devices is set to 51.2 Hz.</p> <p>DATASET FILES</p> <p>The dataset contains the following files:</p> <ul> <li>partX/partXdev1.csv</li> <li>partX/partXdev2.csv</li> <li>partX/partXdev3.csv</li> <li>partX/partXdev4.csv</li> <li>partX/partXdev5.csv</li> </ul> <p>Where X corresponds to the participant ID, and numbers 1-5 to the device IDs indicated in Table 1.</p> <p>Each .csv file has the following format:</p> <ul> <li>Column1: Device ID</li> <li>Column2: accelerometer x</li> <li>Column3: accelerometer y</li> <li>Column4: accelerometer z</li> <li>Column5: gyroscope x</li> <li>Column6: gyroscope y</li> <li>Column7: gyroscope z</li> <li>Column8: magnetometer x</li> <li>Column9: magnetometer y</li> <li>Column10: magnetometer z</li> <li>Column11: Timestamp</li> <li>Column12: Activity Label</li> </ul> <p>Table 1: LOCATIONS</p> <ol> <li>Left Wrist</li> <li>Right Wrist</li> <li>Torso</li> <li>Right Thigh</li> <li>Left Ankle</li> </ol> <p>Table 2: ACTIVITY LABELS</p> <p>(Arrows (->) indicate transitions between activities)</p> <ol> <li>stand</li> <li>sit</li> <li>sit and talk</li> <li>walk</li> <li>walk and talk</li> <li>climb stairs (up/down)</li> <li>climb stairs (up/down) and talk</li> <li>stand -> sit</li> <li>sit -> stand</li> <li>stand -> sit and talk</li> <li>sit and talk -> stand</li> <li>stand -> walk</li> <li>walk -> stand</li> <li>stand -> climb stairs (up/down), stand -> climb stairs (up/down) and talk</li> <li>climb stairs (up/down) -> walk</li> <li>climb stairs (up/down) and talk -> walk and talk</li> </ol>
Multi-task self-supervised learning for wearables - human activity recognition
<p>Datasets used to train and evaluated the self-supervised-learning model</p>
mmWave Radar and RGB-D Camera Sensor Data for Human Activity Recognition
<p>This is a human activity recognition dataset with measurements from both mmWave radar and camera sensor. Meanwhile, we set multiple people scenario to mimic more realistic scenes. The other dataset collected in non-LOS(line-of-sight) environment, you can visit https://zenodo.org/record/7096889#.YynBvuhBwQ8 to get it. The mmWave radar sensors used in our experiments are composed of TI IWR6843ISK-ODS, eradar ESRR(corner radar), eradar EMRR(front radar). We appreciate the support of the eradar company, that provides corner radars and front radars for us, you can visit http://en.eradartech.com/ to get more information. </p>
IMU-based Human Activity Recognition and Payload Estimation for Low-Back Exoskeletons
<p>This repository contains the data for a payload classification algorithm. Data are already split into train, validation and test sets. 12 subjects are included in this dataset (6M+6F).</p> <table> <tbody> <tr> <td><strong>Sensor ID</strong></td> <td><strong>Position</strong></td> </tr> <tr> <td>B6</td> <td>Chest</td> </tr> <tr> <td>3B</td> <td>Right Leg</td> </tr> <tr> <td>BB</td> <td>Right Wrist</td> </tr> <tr> <td>C9</td> <td>Left Leg</td> </tr> <tr> <td>B5</td> <td>Left Wrist</td> </tr> </tbody> </table> <p> </p> <p>This dataset was collected as part of a research project at the Politecnico di Milano and used to train deep-learning algorithms.</p>
DAGHAR: A Benchmark for Domain Adaptation and Generalization in Smartphone-Based Human Activity Recognition
<p>DAGHAR benchmark is a curated dataset collection designed for domain adaptation and domain generalization studies in HAR tasks, using inertial sensors such as accelerometers and gyroscopes, from "A benchmark for domain adaptation and generalization in smartphone-based human activity recognition" work. It features raw inertial sensor data sourced exclusively from smartphones. Six public datasets were selected and standardized in terms of accelerometer units of measurement, sampling rate, gravity component, activity labels, user partitioning, and time window size. This standardization process allows for creating a comprehensive benchmark for evaluating the generalization capabilities of HAR models in cross-dataset scenarios.</p> <p>The benchmark is based on the following datasets:</p> <ul> <li><strong>Ku-HAR</strong>, from "Sikder, N. and Nahid, A.A., 2021. KU-HAR: An open dataset for heterogeneous human activity recognition. Pattern Recognition Letters, 146, pp.46-54", avaliable at <a href="https://data.mendeley.com/datasets/45f952y38r/5">Mendeley</a>. Distributed under CC BY 4.0.</li> <li><strong>MotionSense</strong>, from "Malekzadeh, M., Clegg, R.G., Cavallaro, A. and Haddadi, H., 2019, April. Mobile sensor data anonymization. In Proceedings of the international conference on internet of things design and implementation (pp. 49-58)", available at <a href="https://www.kaggle.com/datasets/malekzadeh/motionsense-dataset" target="_blank" rel="noopener">Kaggle</a>. Distributed under Open Data Commons Open Database License (ODbL) v1.0.</li> <li><strong>RealWorld</strong>, from "Sztyler, T. and Stuckenschmidt, H., 2016, March. On-body localization of wearable devices: An investigation of position-aware activity recognition. In 2016 IEEE international conference on pervasive computing and communications (PerCom) (pp. 1-9). IEEE", available at <a href="https://www.uni-mannheim.de/dws/research/projects/activity-recognition/dataset/dataset-realworld/" target="_blank" rel="noopener">this link</a>. We obtained explicitly permission to distribute a copy of the preprocessed data from the original authors.</li> <li><strong>UCI-HAR</strong>, from "Reyes-Ortiz, J.L., Oneto, L., Samà, A., Parra, X. and Anguita, D., 2016. Transition-aware human activity recognition using smartphones. Neurocomputing, 171, pp.754-767", available at <a href="https://archive.ics.uci.edu/dataset/240/human+activity+recognition+using+smartphones">UCI Repository</a>. Distributed under CC BY 4.0.</li> <li><strong>WISDM</strong>, from "Weiss, G.M., Yoneda, K. and Hayajneh, T., 2019. Smartphone and smartwatch-based biometrics using activities of daily living. Ieee Access, 7, pp.133190-133202", available at <a href="https://archive.ics.uci.edu/dataset/507/wisdm+smartphone+and+smartwatch+activity+and+biometrics+dataset">UCI repository</a>. Distributed under CC BY 4.0.</li> </ul>
Smartphone and smartwatch inertial measurements from heterogeneous subjects for human activity recognition.
<p>This repository contains the dataset and contents described in the <i><strong>"Dataset of inertial measurements of smartphones and smartwatches for human activity recognition"</strong></i> data article.</p><blockquote><p>Matey-Sanz, M., Casteleyn, S., & Granell, C. (2023). Dataset of inertial measurements of smartphones and smartwatches for human activity recognition. <i>Data in Brief</i>, 109809.</p></blockquote>
Dataset: Toothbrushing Data and Analysis of its Potential Use in Human Activity Recognition Applications
<p>This is the dataset that accompanies the paper 'Dataset: Toothbrushing Data and Analysis of its Potential Use in Human Activity Recognition Applications'.</p> <p>In this paper, we describe and analyze a time-series dataset from toothbrushing activity using brush-attached and wearable sensors. The data was collected from 17 participants when they brushed their teeth over one week in 5 different locations. The dataset consists of 62 toothbrushing sessions for each of the brush-attached and wearable sensor approaches, using both electric and manual brushes. The average duration of each session is 2 minutes. One sensor device was attached to the handle of the brush while the other was worn by the participants as a wrist-watch. We collected the data from a 3-axis accelerometer and a 3-axis gyroscope at a 200 Hz sampling rate.</p> <p>Accompanying code can be found at our GitHub repository <a href="https://github.com/icl-mq/toothbrushing-dataset">https://github.com/icl-mq/toothbrushing-dataset</a>. This repository contains example code demonstrating how to process the data along with the file metadata.</p>
Research Data for FiHi: Fusion of inertial and high-resolution acoustic data for privacy-preserving human activity recognition
<h1><strong>Description</strong></h1> <div>This dataset contains information on 20 different activities collected from 15 participants (20-55 years old) using Wit-motion smart inertial sensors and Double Acoustics guitar pickups. Each participant performs these daily activities in an unrestricted environment, with each activity lasting at least 60 seconds and repeated twice.</div> <div> </div> <div>If you use the dataset in an academic work, please cite: </div> <div> </div> <div><code>@ARTICLE{10980212,</code><br><code> author={Yang, Zhe and Zhang, Ying and Li, Yanjun and Huang, Linchong and Hu, Ping and Lin, Yuexiang},</code><br><code> journal={IEEE Transactions on Instrumentation and Measurement}, </code><br><code> title={Fusion of Inertial and High-resolution Acoustic Data for Privacy-Preserving Human Activity Recognition}, </code><br><code> year={2025},</code><br><code> volume={74},</code><br><code> number={},</code><br><code> pages={1-20},</code><br><code> keywords={Human activity recognition;Sensors;Acoustics;Feature extraction;Privacy;Microphones;Biomedical monitoring;Wireless fidelity;Sensor phenomena and characterization;Sensor fusion;Human activities recognition;inertial sensing;Hi-res audio;attention mechanism},</code><br><code> doi={10.1109/TIM.2025.3565250}}</code></div> <h1><strong>DataSet Information</strong></h1> <h2><strong>1.Original_data.zip</strong></h2> <div>The data was annotated by manually reviewing the audio clips and assigning appropriate activity labels. Timestamping the inertial sensor data with the start time of the audio device recorded by the experimenter ensured correct segmentation and synchronization of the inertial and acoustic data. The total data length for</div> <div>all participants is over 10 hours.</div> <h3><strong>(1) </strong><strong>IMU</strong><strong> DATA</strong></h3> <div>The inertial data (accelerometer and gyroscope) is sampled at 100 Hz and transmitted by Bluetooth to the host computer. These reviewed and annotated original samples from 15 participants are placed in separate csv files. The arrangement of information in each csv file is:</div> <div>Col 1-3: 3D-acceleration data (g)</div> <div>Col 4-6: 3D-gyroscope data (°/s)</div> <h3><strong>(2) Audio DATA</strong></h3> <div>The acoustic data are sampled at 192 kHz by a Steinberg sound card and transmitted by USB cable to the host computer. These original samples are placed in separate wav files, with each file name containing all the necessary information regarding the contents of the file.</div> <div><strong>For example:</strong></div> <div>100801_sitting</div> <div>Participant ID (1-4 digits): 1008, Session ID (5-6 digits): 01, Activity ID: sitting.</div> <div> </div> <h2><strong>2、Processed_data.zip</strong></h2> <div>The last two columns of each file are as follows:</div> <ul> <li> <div>participant_id: such as 100101, 100102, 100201 ...... The last two digits are the Session ID, representing the two sessions from the same participant for the same activity.</div> </li> <li> <div>activity_id: Refer to the ACTIVITY SET below</div> </li> </ul> <h3><strong>(1) </strong><strong>IMU</strong><strong> DATA</strong></h3> <div>The original inertial data is individually aligned with the processed Audio data based on their start times, with any excess data rows at the end being trimmed. Then, these inertial data files are augmented with activity_id and participant_id for identification, and consolidated into a single csv file.</div> <div>The continuous motion signal is segmented into sliding windows, each with a duration of 3 seconds and a step size of 3 seconds. Given the IMU’s sampling rate of 100 Hz, each window of inertial data consists of 300 time steps, with 6 channels of information (3 axes each for accelerometer and gyroscope). Consequently, a single inertial sample is represented by a 300 × 6 dimensional matrix.</div> <h3><strong>(2) Audio DATA</strong></h3> <div>The acoustic signals are processed using the Short Time Fourier Transform (STFT) with a window length of 1024 points and an overlap of 256 points, , which generates n linear spaced frequency bins between n kHz frequency range in the frequency domain. The output contains the estimate of the short-term, time-localized frequency patterns. We examine two levels of privacy protection: 8 ∼ 96 kHz for non-speech sound and 20 ∼ 96 kHz for inaudible sound, with 88 and 76 linear spaced frequency bins, respectively.</div> <div>Under a sample rate of 192 kHz for the original acoustic signal, there are (192000 - 256)/(1024 - 256) ≈ 250 frequency features within a second, while each feature has 88 and 76 dimensions for non-speech (8∼96 kHz) and inaudible (20∼96 kHz) feature, respectively.</div> <h1><strong>ACTIVITY </strong><strong>SET</strong></h1> <div>The activityIDs and corresponding activities are listed in the following:</div> <div>0: use microwave</div> <div>1: brush teeth</div> <div>2: browse video</div> <div>3: drink water</div> <div>4: fry</div> <div>5: lie down</div> <div>6: flush</div> <div>7: go downstairs</div> <div>8: go upstairs</div> <div>9: sit</div> <div>10: stand</div> <div>11: manipulate door</div> <div>12: type</div> <div>13: jump</div> <div>14: run</div> <div>15: walk</div> <div>16: wash hands</div> <div>17: write</div> <div>18: operate light</div> <div>19: eat</div>
mmWave Radar and RGB-D Camera Sensor Data for Human Activity Recognition(2)
<p>This is a supplementary dataset, which is linked to https://zenodo.org/record/7088054#.YyVF3ehBwQ8. The dataset is composed of corner radar point cloud data and front radar point cloud data collected from environments existing obstacles between volunteers and sensors. </p>
Data from: Dataset of human medial temporal lobe single neuron activity during declarative memory encoding and recognition
Open the record for dataset details and reuse information.
Smartphone Gyroscope and Accelerometer Dataset for Human Activity Recognition
<p>This dataset is from a study in which we collected smartphone accelerometer and gyroscope data of four subjects while they performed various activities. Two iPhone smartphones were used, with one in the front pants pocket and the other in the back pants pocket. The datasets have been annotated with ground truth activity labels based on video footage. The study procedure is described in our paper: </p> <p>Huang, E.J. and Onnela, J.P. Augmented Movelet Method for Activity Classification Using Smartphone Gyroscope and Accelerometer Data. Sensors 2020, 20, 3706.</p>
Keypoints extracted from custom videos, collected for human activity recognition project
Open the record for dataset details and reuse information.
Spatially-diverse High-dimensional Channel State Information (CSI) based Dataset for Human Activity Recognition (HAR)
<p><strong>Introduction</strong></p> <p>The data collection for this novel channel state information (CSI)-based human activity recognition (HAR) dataset was conducted in a rigorously controlled laboratory environment. The space, measuring 5 meters x 8 meters x 3 meters, served as a well-defined testing ground. Four strategically positioned ESP32 devices formed transceiver pairs in a diagonal network, functioning as both transmitters and receivers, which are separated by a distance of 1.5 meters. The transmitters, powered by external power banks for consistent operation, were mounted on tripods at a height of 1.5 meters on the north and east corners. Their corresponding receivers, connected to laptops via USB for real-time CSI data acquisition, were positioned on the south and west corners.</p> <p>This dataset captures a wider range of activities (including subtle movements) and accounts for variations in body type and environmental conditions. It achieves this by collecting CSI data in a controlled environment using multiple transmitter-receiver pairs positioned at different orientations. This setup captures CSI information across 166 subcarriers used in Wi-Fi Wi-Fi IEEE 802.11n on channel 11, providing a richer and more nuanced view of human movement compared to traditional datasets. This data is expected to lead to the development of more robust and generalizable HAR models with higher accuracy and real-world applicability.</p> <p><strong>Description of Dataset</strong></p> <p>The dataset is housed within a directory named "SHD-HAR-Dataset-main" in the repository. This directory is further divided into "raw" and "amplitude" subdirectories. Data in "raw" and "amplitude" directories is further categorized based on participant orientation relative to the transceivers ("front/side"). It's important to note that the samples are synchronized between the "front" and "side" folders. This is because both devices collected data simultaneously for a particular activity. To summarize, the file "activityX.csv" in the "front" folder was collected at the same time as the corresponding "activityX.csv" file in the "side" folder, where "X" represents a unique identifier for each sample. </p> <p>The "raw" directory stores the unprocessed CSI data for each activity sample. These samples are stored as individual CSV files ("activityX.csv") containing 300-450 packets of CSI data obtained within 5 seconds. Each packet encompasses 25 distinct data fields, resulting in a total of 300-450 x 25 data entries per activity stored within the ".csv" file. </p> <p>The "amplitude" directory contains the processed signal amplitude data for 166 subcarriers. This directory mirrors the structure of the "raw" directory, with subdirectories for "front" and "side" orientations and "activityX.csv" files for each sample. However, the data within these files is transformed into a 300-450 x 166 matrix, representing the extracted signal amplitudes from the original CSI data. This format significantly reduces dimensionality while preserving key information for HAR analysis.</p>
Human Activity Recognition Dataset for Pedestrians with Mobility Disabilities
<p>Human Activity Recognition Dataset for Pedestrians with Mobility Disabilities</p> <p><strong>Note</strong>: The dataset is licensed and shared under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0). To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-sa/4.0/. If you want to refer to this dataset in a publication, please use the reference below: Author names: removed for anonymous review. 2024. Human Activity Recognition Dataset for Pedestrians with Mobility Disabilities Sci. Data 00, 000.</p> <p>1 HAR-PMD Dataset</p> <p>HAR-PMD(Human Activity Recognition for Pedestrians with Mobility Disabilities) dataset consists of smartphone and smartwatch sensor data from six pedestrian activities for people with mobility difficulties: still, walking, crutches, walker, manual wheelchairs, and electric wheelchairs. Sixty participants collected smartphone data, and sixty additional participants collected both smartphone and smartwatch data. Each activity was conducted in both indoor and outdoor environments. Thirteen smartphone sensors and five smartwatch sensors were collected. As a result, the dataset consists of 14,400 minutes of data from 120 participants.</p> <p>For detailed descriptions of activities, environments, sensors, positions, data collection application, and data collection procedure, see METHODS in the paper.</p> <p> </p> <p>2 DATA FORMAT</p> <p>2.1 HAR-PMD Dataset</p> <p>The dataset consists of 120 folders for each individual. Each folder contains metadata, smartphone, and smartwatch data if collected. Smartphone and smartwatch data are stored in comma-separated values (CSV) format files.</p> <p>2.1.1 Metadata</p> <ul> <li><user_id>/device.txt – smartphone model of the participant</li> <li><user_id>/os_version.txt – Android OS version of the smartphone</li> </ul> <p>2.1.2 Smartphone</p> <ul> <li><user_id>/<device>/<user_id>_<activity>_<device>_<environment>.csv – collected smartphone sensor data <ul> <li><user_id>: 1 – 120</li> <li><device>: phone</li> <li><activity>: still, walking, crutches, walker, manual (manual wheelchair), or electric (electric wheelchair)</li> <li><environment>: indoor or outdoor</li> </ul> </li> </ul> <p>2.1.3 Smartwatch</p> <ul> <li><user_id>/<device>/<user_id>_<activity>_<device>_<environment>.csv – collected smartwatch sensor data <ul> <li><user_id>: 61 – 120</li> <li><device>: watch</li> <li><activity>: still, walking, crutches, walker, manual (manual wheelchair), or electric (electric wheelchair)</li> <li><environment>: indoor or outdoor</li> </ul> </li> </ul> <p>2.2 Data Files</p> <p>2.2.1 Smartphone sensor data</p> <p>Table 1 describes each column in a smartphone sensor data CSV file. For more detailed descriptions, see <a href="https://developer.android.com/guide/topics/sensors/sensors_overview">Android Developers Sensors Documents</a>.</p> <p><Table 1: Smartphone sensor data description for each column></p> <table> <tbody> <tr> <th>Columns</th> <th>Description (unit)</th> </tr> <tr> <td>Time</td> <td>Timestamp (s)</td> </tr> <tr> <td>LAccX</td> <td>Acceleration force excluding gravity along the x-axis (m/s²)</td> </tr> <tr> <td>LAccY</td> <td>Acceleration force excluding gravity along the y-axis (m/s²)</td> </tr> <tr> <td>LAccZ</td> <td>Acceleration force excluding gravity along the z-axis (m/s²)</td> </tr> <tr> <td>GyrX</td> <td>Rate of rotation around the x axis (rad/s)</td> </tr> <tr> <td>GyrY</td> <td>Rate of rotation around the y axis (rad/s)</td> </tr> <tr> <td>GyrZ</td> <td>Rate of rotation around the z axis (rad/s)</td> </tr> <tr> <td>MagX</td> <td>Geomagnetic field strength along the x-axis (μT)</td> </tr> <tr> <td>MagY</td> <td>Geomagnetic field strength along the y-axis (μT)</td> </tr> <tr> <td>MagZ</td> <td>Geomagnetic field strength along the z-axis (μT)</td> </tr> <tr> <td>GraX</td> <td>Force of gravity in the x-axis (m/s²)</td> </tr> <tr> <td>GraY</td> <td>Force of gravity in the y-axis (m/s²)</td> </tr> <tr> <td>GraZ</td> <td>Force of gravity in the z-axis (m/s²)</td> </tr> <tr> <td>AccX</td> <td>Acceleration force including gravity along the x-axis (m/s²)</td> </tr> <tr> <td>AccY</td> <td>Acceleration force including gravity along the y-axis (m/s²)</td> </tr> <tr> <td>AccZ</td> <td>Acceleration force including gravity along the z-axis (m/s²)</td> </tr> <tr> <td>Ori_Azimuth</td> <td>Angle around the x-axis (rad)</td> </tr> <tr> <td>Ori_Pitch</td> <td>Angle around the y-axis (rad)</td> </tr> <tr> <td>Ori_Roll</td> <td>Angle around the z-axis (rad)</td> </tr> <tr> <td>RotVec_0</td> <td>Rotation vector component along the x-axis (unitless)</td> </tr> <tr> <td>RotVec_1</td> <td>Rotation vector component along the y-axis (unitless)</td> </tr> <tr> <td>RotVec_2</td> <td>Rotation vector component along the z-axis (unitless)</td> </tr> <tr> <td>RotVec_3</td> <td>Scalar component of the rotation vector (unitless)</td> </tr> <tr> <td>Game_RotVec_0</td> <td>Rotation vector without using geomagnetic filed component along the x-axis (unitless)</td> </tr> <tr> <td>Game_RotVec_1</td> <td>Rotation vector without using geomagnetic filed component along the y-axis (unitless)</td> </tr> <tr> <td>Game_RotVec_2</td> <td>Rotation vector without using geomagnetic filed component along the z-axis (unitless)</td> </tr> <tr> <td>Game_RotVec_3</td> <td>Scalar component without using geomagnetic filed of the rotation vector (unitless)</td> </tr> <tr> <td>Pressure</td> <td>Ambient air pressure (hPa)</td> </tr> <tr> <td>Height</td> <td>Altitude (m)</td> </tr> <tr> <td>Light</td> <td>Illuminance (lx)</td> </tr> <tr> <td>Step</td> <td>Number of steps (steps)</td> </tr> <tr> <td>Proxi</td> <td>Proximity (cm)</td> </tr> </tbody> </table> <p> </p> <p>2.2.2 Smartwatch sensor data</p> <p>Table 2 describes each column in a smartwatch sensor data CSV file. For more detailed descriptions, see <a href="https://developer.android.com/guide/topics/sensors/sensors_overview">Android Developers Sensors Documents</a>.</p> <p><Table 2: Smartwatch sensor data description for each column></p> <table style="width: 42.344%; height: 382.282px;"> <tbody> <tr style="height: 19.5938px;"> <th style="width: 16.3311%; height: 19.5938px;">Columns</th> <th style="width: 83.6689%; height: 19.5938px;">Description (unit)</th> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">Time</td> <td style="width: 83.6689%; height: 19.5938px;">Timestamp (s)</td> </tr> <tr style="height: 39.1875px;"> <td style="width: 16.3311%; height: 39.1875px;">LAccX</td> <td style="width: 83.6689%; height: 39.1875px;">Acceleration force excluding gravity along the x-axis (m/s²)</td> </tr> <tr style="height: 39.1875px;"> <td style="width: 16.3311%; height: 39.1875px;">LAccY</td> <td style="width: 83.6689%; height: 39.1875px;">Acceleration force excluding gravity along the y-axis (m/s²)</td> </tr> <tr style="height: 39.1875px;"> <td style="width: 16.3311%; height: 39.1875px;">LAccZ</td> <td style="width: 83.6689%; height: 39.1875px;">Acceleration force excluding gravity along the z-axis (m/s²)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">GyrX</td> <td style="width: 83.6689%; height: 19.5938px;">Rate of rotation around the x axis (rad/s)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">GyrY</td> <td style="width: 83.6689%; height: 19.5938px;">Rate of rotation around the y axis (rad/s)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">GyrZ</td> <td style="width: 83.6689%; height: 19.5938px;">Rate of rotation around the z axis (rad/s)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">MagX</td> <td style="width: 83.6689%; height: 19.5938px;">Geomagnetic field strength along the x-axis (μT)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">MagY</td> <td style="width: 83.6689%; height: 19.5938px;">Geomagnetic field strength along the y-axis (μT)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">MagZ</td> <td style="width: 83.6689%; height: 19.5938px;">Geomagnetic field strength along the z-axis (μT)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">GraX</td> <td style="width: 83.6689%; height: 19.5938px;">Force of gravity in the x-axis (m/s²)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">GraY</td> <td style="width: 83.6689%; height: 19.5938px;">Force of gravity in the y-axis (m/s²)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">GraZ</td> <td style="width: 83.6689%; height: 19.5938px;">Force of gravity in the z-axis (m/s²)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">AccX</td> <td style="width: 83.6689%; height: 19.5938px;">Acceleration force including gravity along the x-axis (m/s²)</td> </tr> <tr style="height: 10px;"> <td style="width: 16.3311%; height: 10px;">AccY</td> <td style="width: 83.6689%; height: 10px;">Acceleration force including gravity along the y-axis (m/s²)</td> </tr> <tr style="height: 19.5938px;"> <td style="width: 16.3311%; height: 19.5938px;">AccZ</td> <td style="width: 83.6689%; height: 19.5938px;">Acceleration force including gravity along the z-axis (m/s²)</td> </tr> </tbody> </table> <p> </p> <p><span lang="EN-US">3. Benchmark Experiment</span></p> <p><span lang="EN-US">We applied four machine learning models (Decision Tree, Random Forest, XGBoost, and Support Vector Machine) and four deep learning models (Multilayer Perceptron, Convolutional Neural Networks, Long Short-Term Memory (LSTM), and Transformer) to conduct the benchmark experiments. </span></p> <p><span lang="EN-US">We had three validation scenarios: a) Mobility disability; b) Walking aids & wheelchairs; and c) Mobility in detail. Data from all participants and activities were used for each validation scenarios. Each scenario was evaluated using two validation methods: user-depedent (random) evaluation and user-independent evaluation. In the user-dependent (random) evaluation, we randomly shuffled the data and divided the 14,400 min of data into five subsets of 2,880 min each for smartphone sensor data. After shuffling and splitting the data, one subset was allocated for testing while the other four subsets were used for training the model, and this process was repeated for all five subsets. For the combination of smartphone and smartwatch sensors, 7,200 min of data was randomly shuffled and divided the five subsets of 1,440 min data. Each subset was again allocated for testing while the other four subsets were used for training. In the user-independent (UI) evaluation, we used leave-one-group-out 5-fold cross-validation. For smartphone sensor data, 120 participants who comprised the data were divided into five equally distributed subsets of 24 participants each. The models were trained using the data of 96 participants and tested one of the data for the remaining 24 participants. For the combination of smartphone and smartwatch sensors, as the smartwatch data consisted of 60 participants, the data from 48 participants were used for training and the remaining 12 participants were used for testing. This list of participants included in each fold is shown in the below table. In the combination of sensor, we used three sensor combinations: a) linear accelerometer; b) linear accelerometer and gyroscope, and c) linear accelerometer, gyroscope, and magnetometer. Data from all participants and activities were used for each sensor scenarios.</span></p> <p> </p> <table style="width: 67.8639%; height: 136.172px;"> <tbody> <tr style="height: 33.6094px;"> <th style="width: 16.4575%; height: 33.6094px;"> </th> <th style="width: 17.0153%; height: 33.6094px;">Fold 1</th> <th style="width: 16.8759%; height: 33.6094px;">Fold 2</th> <th style="width: 16.5962%; height: 33.6094px;">Fold 3</th> <th style="width: 16.3188%; height: 33.6094px;">Fold 4</th> <th style="width: 16.7364%; height: 33.6094px;">Fold 5</th> </tr> <tr style="height: 47.1875px;"> <td style="width: 16.4575%; height: 47.1875px;">Smartphone only</td> <td style="width: 17.0153%; height: 47.1875px;">1-24</td> <td style="width: 16.8759%; height: 47.1875px;">25-48</td> <td style="width: 16.5962%; height: 47.1875px;">49-72</td> <td style="width: 16.3188%; height: 47.1875px;">73-96</td> <td style="width: 16.7364%; height: 47.1875px;">96-120</td> </tr> <tr style="height: 55.375px;"> <td style="width: 16.4575%; height: 55.375px;">Smartphone & smartwatch</td> <td style="width: 17.0153%; height: 55.375px;">61-72</td> <td style="width: 16.8759%; height: 55.375px;">73-84</td> <td style="width: 16.5962%; height: 55.375px;">85-96</td> <td style="width: 16.3188%; height: 55.375px;">97-108</td> <td style="width: 16.7364%; height: 55.375px;">109-120</td> </tr> </tbody> </table> <p> </p>
mmWave Radar and RGB-D Camera Sensor Data for Human Activity Recognition(2)
<p>This is a supplementary dataset, which is linked to https://zenodo.org/record/7088054#.YyVF3ehBwQ8. The dataset is composed of corner radar point cloud data and front radar point cloud data collected from environments existing obstacles between volunteers and sensors. </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.