Datasets:

PulpBio
/

SilentWear

@@ -1,182 +1,138 @@
 ---
 license: apache-2.0
 ---
-# SilentWear: An Ultra-Low Power Wearable Interface for EMG-Based Silent Speech Recognition
-This repository provides a multi-session surface electromyography (EMG) dataset for vocalized and silent speech recognition, recorded using a wearable neckband interface.
-The dataset is designed to support research in:
-- EMG-based speech decoding
-- Human–machine interaction (HMI)
-- Assistive communication technologies
-- Ultra-low-power wearable AI systems
-The data were collected using **SilentWear**, an unobtrusive, ultra-low-power EMG neckband designed for silent and vocalized speech detection.
-<p align="center" style="white-space: nowrap;">
-  <img src="images/silent_wear_interface.png"
-       alt="SilentWear Device"
-       style="height:300px; display:inline-block; vertical-align:middle;" />
-  <img src="images/signals.png"
-       alt="SilentWear Signals"
-       style="height:300px; display:inline-block; vertical-align:middle;" />
-</p>
 ---
-# Dataset Description
-The dataset includes recordings from:
-- **4 subjects** (3 male, 1 female)
-- **Vocalized** and **silent** speech conditions
-- **8 HMI commands**:
-  *up*, *down*, *left*, *right*, *start*, *stop*, *forward*, *backward*
-  plus a *rest* (no-speech) class
-- **3 recording days** per subject
-- **Multiple sessions, collected over 3 days**, each containing:
-  - 5 vocalized batches.
-  - 5 silent batches
-- Each batch contains *20 repetitions* of each word, plus rest.
-This structure enables evaluation under **multi-day conditions**, supporting research on robustness to electrode repositioning and inter-session variability.
-Further details on the data collection methodology are available at:
 https://arxiv.org/placeholder
 ---
-# Repository Organization
-The repository contains two subfolders:
-### 1️⃣ `data_raw_and_filt`
-This folder contains full-length EMG recordings for each subject,
-condition, session, and batch.
-Each file:
-- Contains raw EMG signals
-- Contains filtered EMG signals (4th-order high-pass at 20 Hz + 50 Hz notch)
-- Is stored in `.h5` format\
-- Uses the HDF5 key `"emg"`
-Directory structure example:
-```text
 data_raw_and_filt/
-└── S01/s
-    └── silent/
-        └── sess_1_batch_1.h5
-        .
-        .
-        └── sess_3_batch_5.h5
     └── vocalized/
-        └── sess_1_batch_1.h5
-        .
-        .
-        └── sess_3_batch_5.h5
-└── S02
-└── S03
-└── S04
-```
-------------------------------------------------------------------------
-#### Example: Loading a File
-``` python
 import pandas as pd
 df = pd.read_hdf("data_raw_and_filt/S01/silent/sess_1_batch_1.h5", key="emg")
-df.head()
-```
-------------------------------------------------------------------------
-#### File Content Structure (`data_raw_and_filt`)
-Each `.h5` file contains:
-```
-  ------------------------------------------------------------------------------
-                    Columns                       Description
-  ---------------- -----------------------        ------------------------------
-  Raw EMG          `Ch_0`--`Ch_15`                Raw data
-  Filtered EMG     `Ch_0_filt`--`Ch_15_filt`      High-pass + notch filtered data
-  Labels           `Label_int`,                   Integer Labels
-                   `Label_str`                    String Labels
-  Session Metadata `session_id`                   Recording session identifier
-  Batch Metadata   `batch_id`                     Batch identifier within session
-  -------------------------------------------------------------------------------
-```
-### 2️⃣ `wins_and_features`
-- Non-overlapping windowed segments
-- Raw and filtered signals
-- Extracted time-frequency features
-These files can be directly used for model training or benchmarking.
 ---
-# Code and Usage
-The dataset is designed to be used in conjunction with the SilentWear repository:
 https://github.com/pulp-bio/silent_wear
-Please refer to the repository `README.md` for:
-- Data loading utilities
-- Preprocessing pipelines
-- Training scripts
-- Evaluation scripts
-The repository creates the files contained in `wins_and_features` folder; these files are then used for model training.
-Alternatively, you may directly use the `data_raw_and_filt` folder to:
-- Build custom dataloaders
-- Train your own architectures
-- Benchmark novel EMG decoding methods
 ---
-#
-# Contributing
-We aim to promote standardized evaluation and fair comparison across models.
-We strongly encourage contributions of trained models and evaluation results to:
-https://github.com/pulp-bio/silent_wear
-Please refer to the repository README for submission guidelines.
 ---
-# Citation
-If you use this dataset, please cite:
-```bibtex
-@online{spacone_silentwear_26,
   author = {Spacone, Giusy and Frey, Sebastian and Pollo, Giovanni and Burrello, Alessio and Pagliari, J. Daniele and Kartsch, Victor and Cossettini, Andrea and Benini, Luca},
-  title = {SilentWear: An Ultra-Low Power Wearable Interface for EMG-Based Silent Speech Recognition},
-  year = {202},
-  url = {https://arxiv.org/placeholder}
 }
-```

 ---
 license: apache-2.0
+language:
+  - en
+tags:
+  - biosignals
+  - emg
+  - silent-speech
+  - speech-recognition
+  - human-machine-interaction
+  - wearable
+  - time-series
+task_categories:
+  - audio-classification
+  - text-classification
+  - signal-processing
+pretty_name: SilentWear EMG Dataset
+dataset_type: other
 ---
+# SilentWear: An Ultra-Low-Power Wearable Interface for EMG-Based Silent Speech Recognition
+This repository provides a multi-session surface electromyography (sEMG) dataset for vocalized and silent speech recognition, recorded using a wearable neckband interface.
+The dataset supports research in:
+- EMG-based speech decoding
+- Human–machine interaction (HMI)
+- Assistive communication technologies
+- Ultra-low-power wearable AI systems
 ---
+## Dataset Summary
+- Subjects: 4 (3 male, 1 female)
+- Conditions: vocalized and silent
+- Commands (8): up, down, left, right, start, stop, forward, backward
+- Additional class: rest (no speech)
+- Days: 3 recording days per subject
+- Sessions:
+  - 5 vocalized batches per session
+  - 5 silent batches per session
+  - 20 repetitions per word per batch
+This structure enables multi-day robustness evaluation (e.g., electrode repositioning and session variability).
+Paper link (replace with final link if needed):
 https://arxiv.org/placeholder
 ---
+## Repository Structure
+### 1) data_raw_and_filt
+Full-length EMG recordings stored as HDF5 (.h5) files using key "emg".
+Each file contains:
+- Raw EMG: Ch_0 – Ch_15
+- Filtered EMG: Ch_0_filt – Ch_15_filt (4th-order high-pass @ 20 Hz + 50 Hz notch)
+- Labels: Label_int, Label_str
+- Session metadata: session_id
+- Batch metadata: batch_id
+Example:
 data_raw_and_filt/
+└── S01/
+    ├── silent/
+    │   ├── sess_1_batch_1.h5
+    │   └── ...
     └── vocalized/
+        ├── sess_1_batch_1.h5
+        └── ...
+└── S02/
+└── S03/
+└── S04/
+Example loading:
 import pandas as pd
 df = pd.read_hdf("data_raw_and_filt/S01/silent/sess_1_batch_1.h5", key="emg")
+print(df.head())
+---
+### 2) wins_and_features
+Contains:
+- Non-overlapping windowed segments
+- Raw and filtered signals
+- Extracted time-frequency features
+These files can be directly used for model training and benchmarking.
 ---
+## Related Code
+Main repository:
 https://github.com/pulp-bio/silent_wear
+Includes:
+- Data loaders
+- Preprocessing pipelines
+- Training scripts
+- Evaluation scripts
 ---
+## Intended Use
+- Silent speech command classification
+- Cross-session robustness studies
+- Low-power wearable EMG decoding research
+---
+## Limitations
+- Small number of subjects (n=4)
+- Single sensing configuration (neckband)
+- Fixed vocabulary command set
+- Cross-user generalization may be limited
 ---
+## Citation
+@online{spacone_silentwear,
   author = {Spacone, Giusy and Frey, Sebastian and Pollo, Giovanni and Burrello, Alessio and Pagliari, J. Daniele and Kartsch, Victor and Cossettini, Andrea and Benini, Luca},
+  title  = {SilentWear: An Ultra-Low Power Wearable Interface for EMG-Based Silent Speech Recognition},
+  year   = {2026},
+  url    = {https://arxiv.org/placeholder}
 }