README.md

---
license: cc-by-nc-4.0
task_categories:
- video-classification
language:
- en
tags:
- medical
- fall-detection
- activity-recognition
- synthetic
pretty_name: 'OmniFall: A Unified Benchmark for Staged-to-Wild Fall Detection'
size_categories:
- 10K<n<100K
---
[![License: CC BY-NC-SA 4.0](https://img.shields.io/badge/License-CC%20BY--NC--SA%204.0-lightgrey.svg)](https://creativecommons.org/licenses/by-nc-sa/4.0/)
</br>
<a href="https://arxiv.org/abs/2505.19889">
  <img src="https://img.shields.io/badge/arXiv-2505.19889-b31b1b.svg?logo=arXiv" alt="arXiv" style="height:20px;">
</a>
</br>
<a href="https://simplexsigil.github.io/omnifall/">
  <img src="https://img.shields.io/badge/Omnifall-Project_Page-2FA7C9.svg" alt="arXiv" style="height:20px;">
</a>

# OmniFall: A Unified Benchmark for Staged-to-Wild Fall Detection

This repository contains annotations, splits, and metadata for OmniFall, a comprehensive benchmark with three components:

- **OF-Staged (OF-Sta):** 8 public staged fall detection datasets (~14h single-view, 112 subjects, 31 camera views)
- **OF-In-the-Wild (OF-ItW):** Curated genuine accident videos from OOPS (~2.7h, 818 videos)
- **OF-Synthetic (OF-Syn):** 12,000 synthetic videos generated with Wan 2.2 (~17h) with controlled demographic diversity

All components share a sixteen-class activity taxonomy. Staged datasets use the core ten classes (0-9), while OF-ItW and OF-Syn use the full sixteen classes (0-15).

[OmniFall: A Unified Staged-to-Wild Benchmark for Human Fall Detection](https://arxiv.org/abs/2505.19889)

Also have a look for additional information on our project page:

[OmniFall Project Page](https://simplexsigil.github.io/omnifall/)

## Datasets

### OF-Staged (8 public datasets)

1. **[CMDFall](https://www.mica.edu.vn/perso/Tran-Thi-Thanh-Hai/CMDFALL.html)** (7h 25m single view) - 50 subjects, 7 synchronized views
2. **[UP Fall](https://sites.google.com/up.edu.mx/har-up/)** (4h 35m) - 17 subjects, 2 synchronized views
3. **[Le2i](https://search-data.ubfc.fr/imvia/FR-13002091000019-2024-04-09_Fall-Detection-Dataset.html)** (47m) - 9 subjects, 6 different rooms
4. **[GMDCSA24](https://github.com/ekramalam/GMDCSA24-A-Dataset-for-Human-Fall-Detection-in-Videos)** (21m) - 4 subjects, 3 rooms
5. **[CAUCAFall](https://data.mendeley.com/datasets/7w7fccy7ky/4)** (16m) - 10 subjects, 1 room
6. **[EDF](https://doi.org/10.5281/zenodo.15494102)** (13m) - 5 subjects, 2 views synchronized
7. **[OCCU](https://doi.org/10.5281/zenodo.15494102)** (14m) - 5 subjects, 2 views not synchronized
8. **[MCFD](https://www.iro.umontreal.ca/~labimage/Dataset/)** (12m) - 1 subject, 8 views

### OF-ItW (In-the-Wild)

9. **[OOPS-Fall](https://oops.cs.columbia.edu/data/)** (2h 39m) - Curated subset of genuine fall accidents from the OOPS dataset, strong variation in subjects and views.

### OF-Syn (Synthetic)

10. **OF-Syn** (16h 53m) - 12,000 synthetic videos generated with Wan 2.2 video diffusion model. Features controlled demographic diversity across age (6 groups), ethnicity (7 OMB categories), body type (4 BMI bands), gender, and environments. Labels include 19 columns: 7 core annotation fields plus 12 demographic and scene metadata fields.

## Structure

The repository is organized as follows:

- `omnifall.py` - Custom HuggingFace dataset builder (handles all configs)
- `labels/` - CSV files containing temporal segment annotations
  - Staged/OOPS labels: 7 columns (`path, label, start, end, subject, cam, dataset`)
  - OF-Syn labels: 19 columns (7 core + 12 demographic/scene metadata)
- `splits/` - Train/validation/test splits
  - `splits/cs/` - Cross-subject splits (staged + OOPS)
  - `splits/cv/` - Cross-view splits (staged + OOPS)
  - `splits/syn/` - Synthetic splits (random, cross_age, cross_ethnicity, cross_bmi)
- `videos/metadata.csv` - OF-Syn video-level metadata (12,000 videos)
- `data_files/syn_frame_wise_labels.tar.zst` - OF-Syn frame-wise HDF5 labels

### Label Format

Each label file in the `labels/` directory follows this format:

```
path,label,start,end,subject,cam,dataset
path/to/clip,class_id,start_time,end_time,subject_id,camera_id,dataset_name
```

Where:
- `path`: Relative path to the video, given the respective dataset root.
- `label`: Class ID (0-15) corresponding to one of sixteen activity classes:
  - 0: walk
  - 1: fall
  - 2: fallen
  - 3: sit_down
  - 4: sitting
  - 5: lie_down
  - 6: lying
  - 7: stand_up
  - 8: standing
  - 9: other
  - 10: kneel_down
  - 11: kneeling
  - 12: squat_down
  - 13: squatting
  - 14: crawl
  - 15: jump
- `start`: Start time of the segment (in seconds)
- `end`: End time of the segment (in seconds)
- `subject`: Subject ID
- `cam`: Camera view ID
- `dataset`: Name of the dataset

**Note on class usage:** Staged datasets (OF-Sta) use only the core ten classes (0-9). OF-ItW and OF-Syn use the full sixteen-class taxonomy. The extended classes (10-15) are infrequent and can be treated as "other" (class 9) for compatibility.

Cam and subject IDs are -1 for OF-ItW and OF-Syn, since these datasets have no consistent camera views or subjects across videos.

**Note on OF-Syn labels:** The `labels/of-syn.csv` file contains 19 columns: the 7 core fields above plus 12 demographic and scene metadata fields (age_group, gender_presentation, monk_skin_tone, race_ethnicity_omb, bmi_band, height_band, environment_category, camera_shot, speed, camera_elevation, camera_azimuth, camera_distance). These additional fields are only available when loading OF-Syn configs.

### Split Format

Split files in the `splits/` directory list the video segments included in each partition. You can use the split paths to filter the label data.:

```
path
path/to/clip
```

## Evaluation Protocols

All configurations are defined in the `omnifall.py` dataset builder and loaded via `load_dataset("simplexsigil2/omnifall", "<config_name>")`.

### Labels (no train/val/test splits)
- `labels` (default): All staged + OOPS labels (52k segments, 7 columns)
- `labels-syn`: OF-Syn labels with demographic metadata (19k segments, 19 columns)
- `metadata-syn`: OF-Syn video-level metadata (12k videos)
- `framewise-syn`: OF-Syn frame-wise HDF5 labels (81 labels per video)

### OF-Staged Configs
- `of-sta-cs`: 8 staged datasets, cross-subject splits
- `of-sta-cv`: 8 staged datasets, cross-view splits

### OF-ItW Config
- `of-itw`: OOPS-Fall in-the-wild genuine accidents

### OF-Syn Configs
- `of-syn`: Fixed randomized 80/10/10 split
- `of-syn-cross-age`: Cross-age split (train: adults, test: children/elderly)
- `of-syn-cross-ethnicity`: Cross-ethnicity split
- `of-syn-cross-bmi`: Cross-BMI split (train: normal/underweight, test: obese)

### Cross-Domain Evaluation
- `of-sta-itw-cs`: Train/val on staged CS, test on OOPS
- `of-sta-itw-cv`: Train/val on staged CV, test on OOPS
- `of-syn-itw`: Train/val on OF-Syn random, test on OOPS

### Aggregate Configs (staged + OOPS combined)
- `cs`: Cross-subject splits for all staged + OOPS
- `cv`: Cross-view splits for all staged + OOPS

### Individual Dataset Configs
- `caucafall`, `cmdfall`, `edf`, `gmdcsa24`, `le2i`, `mcfd`, `occu`, `up_fall`: Individual datasets with cross-subject splits

### Deprecated Config Names

The following old config names still work but emit a deprecation warning:

| Old Name | Use Instead |
|---|---|
| `cs-staged` | `of-sta-cs` |
| `cv-staged` | `of-sta-cv` |
| `cs-staged-wild` | `of-sta-itw-cs` |
| `cv-staged-wild` | `of-sta-itw-cv` |
| `OOPS` | `of-itw` |

## Examples

```python
from datasets import load_dataset

# Load the three main components
staged = load_dataset("simplexsigil2/omnifall", "of-sta-cs")   # 8 staged datasets
itw    = load_dataset("simplexsigil2/omnifall", "of-itw")      # OOPS in-the-wild
syn    = load_dataset("simplexsigil2/omnifall", "of-syn")      # OF-Syn synthetic

# Each returns train/validation/test splits with labels already merged
print(f"Staged train: {len(staged['train'])} segments")
print(f"ItW test: {len(itw['test'])} segments")
print(f"Syn train: {len(syn['train'])} segments")

# Synthetic data includes demographic metadata (19 columns)
example = syn['train'][0]
print(f"Path: {example['path']}, Label: {example['label']}")
print(f"Age: {example['age_group']}, Ethnicity: {example['race_ethnicity_omb']}")

# Cross-domain: train on staged, test on wild
cross = load_dataset("simplexsigil2/omnifall", "of-sta-itw-cs")
print(f"Train on staged: {len(cross['train'])} segments")
print(f"Test on wild: {len(cross['test'])} segments")

# All labels without splits (for custom splitting)
labels = load_dataset("simplexsigil2/omnifall", "labels")["train"]
syn_labels = load_dataset("simplexsigil2/omnifall", "labels-syn")["train"]

# Synthetic frame-wise labels (81 labels per video)
framewise = load_dataset("simplexsigil2/omnifall", "of-syn", framewise=True)
print(f"Frame labels: {framewise['train'][0]['frame_labels'][:5]}...")  # [1, 1, 1, ...]
```

## Label definitions

In this section we provide additional information about the labelling process to provide as much transparency as possible.

The benchmark uses a sixteen-class taxonomy. Staged datasets use classes 0-9, while OOPS-Fall uses the full range 0-15.

### Core Classes (0-9, all datasets)

- **`0|walk`** - Move around, including jogging and running and "drunk walking", but only if it is not part of some special exercise like pulling your knees up. Not when pushing a large object like a chair, but included carrying something small like an apple.
- **`1|fall`** - The act of falling (from any previous state). Includes falling on a bed, if the process is not a controlled lying down with arms as support.
- **`2|fallen`** - Being on the ground or a mattress after a fall.
- **`3|sit_down`** - Sitting down on bed or chair or ground.
- **`4|sitting`** - Sitting on bed or chair or ground.
- **`5|lie_down`** - Lying down intentionally (in contrast to a fall) on ground or bed.
- **`6|lying`** - Being in a lying position (in bed or on the ground) after intentionally getting into that position.
- **`7|stand_up`** - Standing up from a fallen state, from lying or sitting. Includes getting from lying position into sitting position.
- **`8|standing`** - Standing around without walking.
- **`9|other`** - Any other activity, including e.g. walking while pushing an object like a chair.

### Extended Classes (10-15, OF-ItW and OF-Syn)

These classes capture additional activities observed in genuine accident and synthetic videos. They are infrequent and do not occur in staged datasets.

- **`10|kneel_down`** - Transitioning from standing or another posture to a kneeling position.
- **`11|kneeling`** - Being in a kneeling position (static posture).
- **`12|squat_down`** - Transitioning to a squatting position.
- **`13|squatting`** - Being in a squatting position (static posture).
- **`14|crawl`** - Crawling on hands and knees or similar locomotion on the ground.
- **`15|jump`** - Jumping action, including vertical jumps and jumps from elevated positions.

## Motion Types

There are two types of motions, **dynamic** ones like `walk` or `stand_up` and **static** ones like `fallen`, `sitting`, `lying`.

**Generally we annotate dynamic motions as soon as the first frame appears which belongs to that action**.

Let's say we see a person `walk`, then `fall`. The first frame which indicates a motion which does seem to be different to `walk` is the start of `fall`. Sometimes it is necessary to have a look at a person walking to learn when the motion begins to change to something else.

**For static motions the label begins with the first frame where the person comes to a resting state.** For `sit_down`, the label ends when the person reaches a state where it is no longer adjusting its body position but comes to a rest. `fall` ends when the person is no longer moving caused by the inertia of the fall. The following `fallen` might contain movement on the ground, but no movement which belongs to `fall` or `stand_up`.

## Label Sequences

There are some natural sequences of labels like `fall`, `fallen` and `stand_up`. However, it is not always the case that these appear together. Sometimes the person might directly stand up again without any time at rest on the ground, in this case there is no `fallen` segment. Likewise sometimes there is no sitting segment after sitting down (e.g. if the person immediately stands up again).

Lying down can be on a bed or on the ground, it is intentional in contrast to `fall`. There are falls which follow `sit_down` or `lie_down` if it is from a chair or from a bed.

When a person is lying in a bed and getting up to sit in the bed we label this as `stand_up`, even if the person is still sitting in the bed. A sequence could then be `lying`, `stand_up`, `sitting`, `stand_up` to describe a person which first lies down, then gets into a sitting position, waits a little, then gets fully up.

Sometimes it is not perfectly clear if it is `sit_down` followed by `lie_down` or simply `lie_down`. This depends on there being a moment of rest or not. If the person spends a short amount of time in the sitting position it is the former, if the person directly goes from `standing` over `sit_down` to `lie_down` without rest it is only labeled `lie_down`. Similar thoughts apply to `stand_up`.

In the video at the bottom of this page we show how we leveraged VGG VIA to perform the annotations. Note, that we pre-load the original dataset labels as visual aid, but relabel the whole video with our label definitions. In the video below, CMDFall is shown, which already provides relatively detailed labels, the original labels of other datasets are more sparse. Additionally, even CMDFall does not label all frames, but only specific segments, while our labels cover mostly every frame.

The blurred out regions were added by us in post-processing to protect the subjects privacy on this page, they are not part of the original videos.

## Citation

If you use OmniFall in your research, please cite our paper (will be updated soon) as well as all sub-dataset papers:

```bibtex
@misc{omnifall,
      title={OmniFall: From Staged Through Synthetic to Wild, A Unified Multi-Domain Dataset for Robust Fall Detection}, 
      author={David Schneider and Zdravko Marinov and Rafael Baur and Zeyun Zhong and Rodi Düger and Rainer Stiefelhagen},
      year={2025},
      eprint={2505.19889},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2505.19889}, 
},

@inproceedings{omnifall_cmdfall,
  title={A multi-modal multi-view dataset for human fall analysis and preliminary investigation on modality},
  author={Tran, Thanh-Hai and Le, Thi-Lan and Pham, Dinh-Tan and Hoang, Van-Nam and Khong, Van-Minh and Tran, Quoc-Toan and Nguyen, Thai-Son and Pham, Cuong},
  booktitle={2018 24th International Conference on Pattern Recognition (ICPR)},
  pages={1947--1952},
  year={2018},
  organization={IEEE}
},

@article{omnifall_up-fall,
  title={UP-fall detection dataset: A multimodal approach},
  author={Mart{\'\i}nez-Villase{\~n}or, Lourdes and Ponce, Hiram and Brieva, Jorge and Moya-Albor, Ernesto and N{\'u}{\~n}ez-Mart{\'\i}nez, Jos{\'e} and Pe{\~n}afort-Asturiano, Carlos},
  journal={Sensors},
  volume={19},
  number={9},
  pages={1988},
  year={2019},
  publisher={MDPI}
},

@article{omnifall_le2i,
  title={Optimized spatio-temporal descriptors for real-time fall detection: comparison of support vector machine and Adaboost-based classification},
  author={Charfi, Imen and Miteran, Johel and Dubois, Julien and Atri, Mohamed and Tourki, Rached},
  journal={Journal of Electronic Imaging},
  volume={22},
  number={4},
  pages={041106--041106},
  year={2013},
  publisher={Society of Photo-Optical Instrumentation Engineers}
},

@article{omnifall_gmdcsa,
  title={GMDCSA-24: A dataset for human fall detection in videos},
  author={Alam, Ekram and Sufian, Abu and Dutta, Paramartha and Leo, Marco and Hameed, Ibrahim A},
  journal={Data in Brief},
  volume={57},
  pages={110892},
  year={2024},
  publisher={Elsevier}
},

@article{omnifall_cauca,
  title={Dataset CAUCAFall},
  author={Eraso, Jose Camilo and Mu{\~n}oz, Elena and Mu{\~n}oz, Mariela and Pinto, Jesus},
  journal={Mendeley Data},
  volume={4},
  year={2022}
},

@inproceedings{omnifall_edf_occu,
  title={Evaluating depth-based computer vision methods for fall detection under occlusions},
  author={Zhang, Zhong and Conly, Christopher and Athitsos, Vassilis},
  booktitle={International symposium on visual computing},
  pages={196--207},
  year={2014},
  organization={Springer}
},

@article{omnifall_mcfd,
  title={Multiple cameras fall dataset},
  author={Auvinet, Edouard and Rougier, Caroline and Meunier, Jean and St-Arnaud, Alain and Rousseau, Jacqueline},
  journal={DIRO-Universit{\'e} de Montr{\'e}al, Tech. Rep},
  volume={1350},
  pages={24},
  year={2010}
},

@inproceedings{omnifall_oops,
  title={Oops! predicting unintentional action in video},
  author={Epstein, Dave and Chen, Boyuan and Vondrick, Carl},
  booktitle={Proceedings of the IEEE/CVF conference on computer vision and pattern recognition},
  pages={919--929},
  year={2020}
}
```

## License

The annotations and split definitions in this repository are released under [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License](https://creativecommons.org/licenses/by-nc-sa/4.0/).

The original video data belongs to their respective owners and should be obtained from the original sources.

## Contact

For questions about the dataset, please contact [david.schneider@kit.edu].


## How we used VGG VIA to densely annotate videos for Omnifall
We provide this video to shortly demonstrate how the annotation process was conducted, increasing transparency. Note, that the CMDFall dataset already provides realtively detailed labels, this is not the case for many of the other datasets.

<video src="https://huggingface.co/datasets/simplexsigil2/omnifall/resolve/main/annotation_remarks.webm" controls />