3DO Dataset | On the Generalization of WiFi-based Person-centric Sensing in Through-Wall Scenarios
<p><strong>On the Generalization of WiFi-based Person-centric Sensing in Through-Wall Scenarios</strong></p>
<p>This repository contains the <strong>3DO dataset</strong> proposed in <a href="https://doi.org/10.1007/978-3-031-78354-8_13">[1]</a>.</p>
<p><strong>PyTroch Dataloader</strong></p>
<p>A minimal PyTorch dataloader for the 3DO dataset is provided at: <a href="https://github.com/StrohmayerJ/3DO/tree/main">https://github.com/StrohmayerJ/3DO</a></p>
<p><strong>Dataset Description</strong></p>
<p>The 3DO dataset comprises 42 five-minute recordings (~1.25M WiFi packets) of three human activities performed by a single person, captured in a WiFi through-wall sensing scenario over three consecutive days. Each WiFi packet is annotated with a 3D trajectory label and a class label for the activities: no person/background (0), walking (1), sitting (2), and lying (3). (<strong>Note:</strong> The labels returned in our dataloader example are walking (0), sitting (1), and lying (2), because background sequences are not used.)</p>
<p>The directories <code>3DO/d1/</code>, <code>3DO/d2/</code>, and <code>3DO/d3/</code> contain the sequences from days 1, 2, and 3, respectively. Furthermore, each sequence directory (e.g., <code>3DO/d1/w1/</code>) contains a <code>csiposreg.csv</code> file storing the raw WiFi packet time series and a <code>csiposreg_complex.npy</code> cache file, which stores the complex Channel State Information (CSI) of the WiFi packet time series. (If missing, <code>csiposreg_complex.npy</code> is automatically generated by the provided dataloader.)</p>
<p>Dataset Structure:</p>
<p>/3DO</p>
<p>├── d1 <em><-- day 1 subdirectory</em></p>
<p> └── w1 <em><-- sequence subdirectory</em></p>
<p> └── csiposreg.csv <em><-- raw WiFi packet time series</em></p>
<p> └── csiposreg_complex.npy <em><-- CSI time series cache</em></p>
<p>├── d2 <-- day 2 subdirectory</p>
<p>├── d3 <-- day 3 subdirectory</p>
<p> </p>
<p>In [1], we use the following training, validation, and test split:</p>
<table>
<tbody>
<tr>
<td><strong>Subset</strong></td>
<td><strong>Day</strong></td>
<td><strong>Sequences </strong></td>
</tr>
<tr>
<td>Train</td>
<td>1</td>
<td>w1, w2, w3, s1, s2, s3, l1, l2, l3</td>
</tr>
<tr>
<td>Val</td>
<td>1</td>
<td>w4, s4, l4</td>
</tr>
<tr>
<td>Test</td>
<td>1</td>
<td>w5 , s5, l5</td>
</tr>
<tr>
<td>Test</td>
<td>2</td>
<td>w1, w2, w3, w4, w5, s1, s2, s3, s4, s5, l1, l2, l3, l4, l5</td>
</tr>
<tr>
<td>Test</td>
<td>3</td>
<td>w1, w2, w4, w5, s1, s2, s3, s4, s5, l1, l2, l4</td>
</tr>
</tbody>
</table>
<p><em>w = walking, s = sitting and l= lying</em></p>
<p><strong>Note: </strong>On each day, we additionally recorded three ten-minute background sequences (b1, b2, b3), which are provided as well.</p>
<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 our paper <a href="https://doi.org/10.1007/978-3-031-78354-8_13" target="_blank" rel="noopener">[1]</a>.</p>
<p><a href="https://doi.org/10.1007/978-3-031-78354-8_13" target="_blank" rel="noopener">[1]</a> Strohmayer, J., Kampel, M. (2025). On the Generalization of WiFi-Based Person-Centric Sensing in Through-Wall Scenarios. In: Pattern Recognition. ICPR 2024. Lecture Notes in Computer Science, vol 15315. Springer, Cham. <a href="https://doi.org/10.1007/978-3-031-78354-8_13" target="_blank" rel="noopener">https://doi.org/10.1007/978-3-031-78354-8_13</a></p>
<p>BibTeX citation:</p>
<pre>@inproceedings{strohmayerOn2025,
author="Strohmayer, Julian and Kampel, Martin",<br> title="On the Generalization of WiFi-Based Person-Centric Sensing in Through-Wall Scenarios",<br> booktitle="Pattern Recognition",<br> year="2025",<br> publisher="Springer Nature Switzerland",<br> address="Cham",<br> pages="194--211",<br> isbn="978-3-031-78354-8"
}</pre>
opencc-by-4.0Nov 2024View details →
