Datasets:
Tasks:
Audio Classification
Modalities:
Audio
Formats:
soundfolder
Languages:
English
Size:
< 1K
Tags:
footstep-detection
footstep-audio
sound-event-detection
audio-classification
acoustic-recognition
walking-sounds
License:
Update README.md
Browse files
README.md
CHANGED
|
@@ -33,17 +33,16 @@ modality:
|
|
| 33 |
---
|
| 34 |
# Footstep Detection Dataset — 50 Hours of Real Footstep Audio
|
| 35 |
|
| 36 |
-
**50 hours of real footstep audio recordings** for training footstep detection, sound event detection, and audio classification models. 166 manually verified files captured in natural indoor and outdoor conditions, with per-file metadata on surface, footwear, location, and background noise
|
| 37 |
|
| 38 |
## Contact us and share your feedback — receive additional samples for free! 😊
|
| 39 |
|
| 40 |
## Key Highlights
|
| 41 |
|
| 42 |
- **50 hours** of real-world footstep audio
|
| 43 |
-
- **166 manually verified files** — every recording reviewed for clear footstep audibility
|
| 44 |
- **Indoor + outdoor** capture conditions
|
| 45 |
-
- **
|
| 46 |
-
- **
|
| 47 |
- **No synthetic audio, no augmentation, no AI-generated content**
|
| 48 |
- Smartphone-first recordings (matches real deployment conditions)
|
| 49 |
|
|
@@ -56,46 +55,13 @@ modality:
|
|
| 56 |
- **Activity recognition** — elderly care, fall detection, ambient assisted living
|
| 57 |
- **Foley generation** — training V2A models for walking sounds in games and animation
|
| 58 |
|
| 59 |
-
## Dataset Structure
|
| 60 |
-
|
| 61 |
-
```
|
| 62 |
-
footstep-detection-dataset/
|
| 63 |
-
├── audio/
|
| 64 |
-
│ ├── rec_001.wav
|
| 65 |
-
│ ├── rec_002.wav
|
| 66 |
-
│ └── ... (158 WAV + 8 M4A files)
|
| 67 |
-
├── metadata.csv
|
| 68 |
-
└── README.md
|
| 69 |
-
```
|
| 70 |
-
|
| 71 |
-
### metadata.csv schema
|
| 72 |
-
|
| 73 |
-
| Field | Type | Values |
|
| 74 |
-
|-------|------|--------|
|
| 75 |
-
| `file_id` | string | unique recording ID |
|
| 76 |
-
| `filename` | string | path to audio file |
|
| 77 |
-
| `duration_sec` | float | 10–100 seconds |
|
| 78 |
-
| `sample_rate` | int | 48000 (majority), 44100, 16000 |
|
| 79 |
-
| `channels` | int | 1 (mono) or 2 (stereo) |
|
| 80 |
-
| `format` | string | wav, m4a |
|
| 81 |
-
| `surface` | string | wood_laminate, tile, carpet, concrete_asphalt, stairs, other |
|
| 82 |
-
| `footwear` | string | barefoot, slippers, sandals, sneakers, dress_shoes_boots, other |
|
| 83 |
-
| `location` | string | indoor, outdoor |
|
| 84 |
-
| `noise_level` | string | low, medium, high |
|
| 85 |
-
| `device_class` | string | smartphone, laptop, tablet |
|
| 86 |
-
|
| 87 |
## Dataset Statistics
|
| 88 |
|
| 89 |
| Metric | Value |
|
| 90 |
|--------|-------|
|
| 91 |
| Total duration | 50 hours |
|
| 92 |
-
| Total files | 166 |
|
| 93 |
-
| WAV files | 158 |
|
| 94 |
-
| M4A files | 8 |
|
| 95 |
| File duration range | 10–100 sec |
|
| 96 |
| Sample rates | 48 kHz / 44.1 kHz / 16 kHz |
|
| 97 |
-
| Surface categories | 6 |
|
| 98 |
-
| Footwear categories | 6 |
|
| 99 |
| Capture conditions | indoor + outdoor |
|
| 100 |
|
| 101 |
## How This Compares to Academic Footstep Audio Datasets
|
|
@@ -109,64 +75,19 @@ footstep-detection-dataset/
|
|
| 109 |
| ESC-50 | <0.1h equivalent | 40 samples | None (label only) |
|
| 110 |
| PURE | 14 minutes | 14 samples | 5 subjects |
|
| 111 |
|
| 112 |
-
|
| 113 |
-
|
| 114 |
-
```python
|
| 115 |
-
from datasets import load_dataset
|
| 116 |
-
|
| 117 |
-
dataset = load_dataset("AxonData/footstep-detection-dataset")
|
| 118 |
-
print(dataset)
|
| 119 |
-
|
| 120 |
-
sample = dataset["train"][0]
|
| 121 |
-
print(sample["audio"]) # audio array + sampling_rate
|
| 122 |
-
print(sample["surface"]) # e.g. "wood_laminate"
|
| 123 |
-
print(sample["footwear"]) # e.g. "sneakers"
|
| 124 |
-
print(sample["noise_level"]) # e.g. "low"
|
| 125 |
-
```
|
| 126 |
-
|
| 127 |
-
## Quick Start — PyTorch DataLoader
|
| 128 |
-
|
| 129 |
-
```python
|
| 130 |
-
import torch
|
| 131 |
-
import torchaudio
|
| 132 |
-
from datasets import load_dataset
|
| 133 |
-
|
| 134 |
-
ds = load_dataset("AxonData/footstep-detection-dataset", split="train")
|
| 135 |
-
|
| 136 |
-
def collate(batch):
|
| 137 |
-
waveforms = [torch.tensor(item["audio"]["array"]) for item in batch]
|
| 138 |
-
labels = [item["surface"] for item in batch]
|
| 139 |
-
return waveforms, labels
|
| 140 |
-
|
| 141 |
-
loader = torch.utils.data.DataLoader(ds, batch_size=8, collate_fn=collate)
|
| 142 |
-
```
|
| 143 |
-
|
| 144 |
-
## Sample vs Full Version
|
| 145 |
-
|
| 146 |
-
This HuggingFace repository contains a **sample subset** for evaluation. The full 50-hour dataset is licensed for commercial use through Axon Labs.
|
| 147 |
-
|
| 148 |
-
**Full version of dataset is available for commercial usage — leave a request on our website [Axonlabs](https://axonlab.ai/dataset/footsteps-audio-dataset/?utm_source=hugging-face&utm_medium=cpc&utm_campaign=footstep&utm_content=readme) to purchase the dataset 💰**
|
| 149 |
|
| 150 |
## What Makes This Dataset Unique
|
| 151 |
|
| 152 |
-
- **Largest footstep audio corpus available commercially**
|
| 153 |
-
- **Manually verified, not scraped**
|
| 154 |
-
- **Real smartphone recordings**
|
| 155 |
-
- **Structured metadata
|
| 156 |
-
- **Backed by a biometric AI specialist** — Axon Labs builds datasets used by 21% of iBeta 2025 certified companies
|
| 157 |
-
|
| 158 |
-
## Two Dataset Versions Available
|
| 159 |
|
| 160 |
-
-
|
| 161 |
-
- **Full Version** — 50 hours of audio with complete metadata, licensed for commercial training
|
| 162 |
-
|
| 163 |
-
[Contact us](https://axonlab.ai/dataset/footsteps-audio-dataset/) to choose the version that fits your project.
|
| 164 |
|
| 165 |
## FAQ
|
| 166 |
|
| 167 |
-
**Q: What's the largest publicly available footstep audio dataset?**
|
| 168 |
-
This one — 50 hours of curated recordings, 3–5× larger than AFPILD (10h) or AFPID-II (14h), which are the most cited academic benchmarks in the field. Sound event datasets like FSD50K and ESC-50 contain footsteps only as a small subset (under 1,000 samples).
|
| 169 |
-
|
| 170 |
**Q: Can I use this dataset for footstep biometrics / acoustic person identification?**
|
| 171 |
Yes. The dataset is well-suited for footstep biometrics research, especially as a pre-training corpus. For per-subject identification tasks, we can collect additional per-subject sessions on request through our custom data collection service.
|
| 172 |
|
|
@@ -176,21 +97,6 @@ Yes. The dataset is well-suited for footstep biometrics research, especially as
|
|
| 176 |
**Q: Is the data ethically collected?**
|
| 177 |
Yes. All recordings were captured with explicit participant consent and processed in accordance with GDPR. Full documentation of consent and provenance is available for the commercial version.
|
| 178 |
|
| 179 |
-
## Citation
|
| 180 |
-
|
| 181 |
-
If you use this dataset in your research, please cite:
|
| 182 |
-
|
| 183 |
-
```bibtex
|
| 184 |
-
@misc{axonlabs2026footstep,
|
| 185 |
-
title = {Footstep Detection Audio Dataset},
|
| 186 |
-
author = {Axon Labs},
|
| 187 |
-
year = {2026},
|
| 188 |
-
url = {https://axonlab.ai/dataset/footsteps-audio-dataset/}
|
| 189 |
-
}
|
| 190 |
-
```
|
| 191 |
-
|
| 192 |
-
---
|
| 193 |
-
|
| 194 |
**keywords**: footstep audio dataset, footstep sound dataset, footstep detection dataset, sound event detection, audio classification dataset, acoustic person identification, footstep biometrics, walking surface classification, foley dataset, environmental sound dataset, real-world audio dataset, smart home audio, activity recognition
|
| 195 |
|
| 196 |
-
Visit us at [**Axonlabs**](https://axonlab.ai/?utm_source=hugging-face&utm_medium=cpc&utm_campaign=
|
|
|
|
| 33 |
---
|
| 34 |
# Footstep Detection Dataset — 50 Hours of Real Footstep Audio
|
| 35 |
|
| 36 |
+
**50 hours of real footstep audio recordings** for training footstep detection, sound event detection, and audio classification models. 166 manually verified files captured in natural indoor and outdoor conditions, with per-file metadata on surface, footwear, location, and background noise
|
| 37 |
|
| 38 |
## Contact us and share your feedback — receive additional samples for free! 😊
|
| 39 |
|
| 40 |
## Key Highlights
|
| 41 |
|
| 42 |
- **50 hours** of real-world footstep audio
|
|
|
|
| 43 |
- **Indoor + outdoor** capture conditions
|
| 44 |
+
- **Different surface categories** annotated per file
|
| 45 |
+
- **Different footwear categories** annotated per file
|
| 46 |
- **No synthetic audio, no augmentation, no AI-generated content**
|
| 47 |
- Smartphone-first recordings (matches real deployment conditions)
|
| 48 |
|
|
|
|
| 55 |
- **Activity recognition** — elderly care, fall detection, ambient assisted living
|
| 56 |
- **Foley generation** — training V2A models for walking sounds in games and animation
|
| 57 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
## Dataset Statistics
|
| 59 |
|
| 60 |
| Metric | Value |
|
| 61 |
|--------|-------|
|
| 62 |
| Total duration | 50 hours |
|
|
|
|
|
|
|
|
|
|
| 63 |
| File duration range | 10–100 sec |
|
| 64 |
| Sample rates | 48 kHz / 44.1 kHz / 16 kHz |
|
|
|
|
|
|
|
| 65 |
| Capture conditions | indoor + outdoor |
|
| 66 |
|
| 67 |
## How This Compares to Academic Footstep Audio Datasets
|
|
|
|
| 75 |
| ESC-50 | <0.1h equivalent | 40 samples | None (label only) |
|
| 76 |
| PURE | 14 minutes | 14 samples | 5 subjects |
|
| 77 |
|
| 78 |
+
**Full version of dataset is available for commercial usage — leave a request on our website [Axonlabs](https://axonlab.ai/?utm_source=hugging-face&utm_medium=cpc&utm_campaign=profile&utm_content=profile_link) to purchase the dataset 💰**
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 79 |
|
| 80 |
## What Makes This Dataset Unique
|
| 81 |
|
| 82 |
+
- **Largest footstep audio corpus available commercially** - 3–5× larger than the most cited academic alternatives
|
| 83 |
+
- **Manually verified, not scraped** - every file reviewed for clear footstep audibility
|
| 84 |
+
- **Real smartphone recordings** - matches deployment conditions for smart speakers, phones, wearables
|
| 85 |
+
- **Structured metadata** - supports filtered training and multi-task learning
|
|
|
|
|
|
|
|
|
|
| 86 |
|
| 87 |
+
[Contact us](https://axonlab.ai/?utm_source=hugging-face&utm_medium=cpc&utm_campaign=profile&utm_content=profile_link) to choose the version that fits your project.
|
|
|
|
|
|
|
|
|
|
| 88 |
|
| 89 |
## FAQ
|
| 90 |
|
|
|
|
|
|
|
|
|
|
| 91 |
**Q: Can I use this dataset for footstep biometrics / acoustic person identification?**
|
| 92 |
Yes. The dataset is well-suited for footstep biometrics research, especially as a pre-training corpus. For per-subject identification tasks, we can collect additional per-subject sessions on request through our custom data collection service.
|
| 93 |
|
|
|
|
| 97 |
**Q: Is the data ethically collected?**
|
| 98 |
Yes. All recordings were captured with explicit participant consent and processed in accordance with GDPR. Full documentation of consent and provenance is available for the commercial version.
|
| 99 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 100 |
**keywords**: footstep audio dataset, footstep sound dataset, footstep detection dataset, sound event detection, audio classification dataset, acoustic person identification, footstep biometrics, walking surface classification, foley dataset, environmental sound dataset, real-world audio dataset, smart home audio, activity recognition
|
| 101 |
|
| 102 |
+
Visit us at [**Axonlabs**](https://axonlab.ai/?utm_source=hugging-face&utm_medium=cpc&utm_campaign=profile&utm_content=profile_link) to request a full version of the dataset for commercial usage
|