Upload folder using huggingface_hub

README.md

@@ -1,3 +1,194 @@
----
-license: cc-by-nc-4.0
----
+# OmniFall: A Unified Benchmark for Staged-to-Wild Fall Detection
+
+[![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/)
+
+This repository contains the annotation and split definitions for OmniFall, a comprehensive benchmark that unifies eight public indoor fall datasets under a consistent ten-class annotation scheme, complemented by the OOPS-Fall benchmark of genuine accidents captured in the wild.
+
+**Paper:** [OmniFall: A Unified Staged-to-Wild Benchmark for Human Fall Detection]()
+
+## Overview
+
+Falls are the leading cause of fatal injuries among older adults worldwide. While the mechanical event of falling lasts only a fraction of a second, the critical health risk often comes from the ensuing "fallen" state—when a person remains on the ground, potentially injured and unable to call for help.
+
+OmniFall addresses three critical limitations in current fall detection research:
+
+1. **Unified Taxonomy:** Rather than binary fall/no-fall classification, we provide a consistent ten-class scheme across datasets that distinguishes transient actions (fall, sit down, lie down, stand up) from their static outcomes (fallen, sitting, lying, standing).
+
+2. **Combined Benchmark:** We unify eight public datasets (14+ hours of video, 112 subjects, 31 camera views) into a single benchmark with standardized train/val/test splits.
+
+3. **In-the-Wild Evaluation:** We include OOPS-Fall, curated from genuine accident videos of the OOPS dataset to test generalization to real-world conditions.
+
+## Datasets
+
+This benchmark includes annotations for the following datasets:
+
+1. **CMDFall** (7h 7m) - Multi-subject, multi-view dataset
+2. **UP Fall** (4h 35m) - Multiple subjects with varied camera views
+3. **Le2i** (47m) - Multiple room settings
+4. **GMDCSA24** (21m) - Varied lighting conditions
+5. **CAUCAFall** (16m) - Includes day/night recordings
+6. **OCCU** (14m) - Depth camera recordings
+7. **EDF** (13m) - Depth camera recordings
+8. **MCFD** (12m) - Single subject, eight camera views
+9. **OOPS-Fall** - Curated subset of genuine accidents from the OOPS dataset
+
+## Structure
+
+The repository is organized as follows:
+
+- `labels/` - CSV files containing frame-level annotations for each dataset as well as label2id.csv
+- `splits/` - Train/validation/test splits for cross-subject (CS) and cross-view (CV) evaluation
+  - `splits/cs/` - Cross-subject splits, where training, validation, and test sets contain different subjects
+  - `splits/cv/` - Cross-view splits, where training, validation, and test sets contain different camera views
+- `dataset.yaml` - Hugging Face datasets configuration with multiple evaluation protocols
+
+### Label Format
+
+Each label file in the `labels/` directory follows this format:
+
+```
+path,label,start,end,subject,cam
+path/to/clip,class_id,start_time,end_time,subject_id,camera_id
+```
+
+Where:
+- `path`: Relative path to the video, given the respective dataset root.
+- `label`: Class ID (0-9) corresponding to one of the ten 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
+- `start`: Start time of the segment (in seconds)
+- `end`: End time of the segment (in seconds)
+- `subject`: Subject ID
+- `cam`: Camera view ID
+
+For OOPS-Fall, only fall segments and non-fall segments are labeled; non-falls are labels as "other", independent of the underlying content, as long as it is not a fall.
+
+### Split Format
+
+Split files in the `splits/` directory list the video segments included in each partition:
+
+```
+path
+path/to/clip
+```
+
+## Evaluation Protocols
+
+We provide multiple evaluation configurations via the `dataset.yaml` file:
+
+### Basic Configurations
+- `default`: Access to all dataset labels
+- `cs`: Cross-subject splits for all datasets
+- `cv`: Cross-view splits for all datasets
+
+### Individual Dataset Configurations
+- `caucafall`, `cmdfall`, `edf`, `gmdcsa24`, `le2i`, `mcfd`, `occu`, `up_fall`, `OOPS`: 
+  Access to individual datasets with their respective cross-subject splits
+
+### Multi-Dataset Evaluation Protocols
+- `cs-staged`: Cross-subject splits combined across all staged datasets
+- `cv-staged`: Cross-view splits combined across all staged datasets
+- `cs-staged-wild`: Train and validate on staged datasets with cross-subject splits, test on OOPS-Fall
+- `cv-staged-wild`: Train and validate on staged datasets with cross-view splits, test on OOPS-Fall
+
+## Usage
+
+To use this dataset with the Hugging Face datasets library:
+
+```python
+from datasets import load_dataset
+
+# Load the entire dataset with default configuration
+dataset = load_dataset("omnifall")
+
+# Use cross-subject (CS) evaluation protocol
+cs_dataset = load_dataset("omnifall", "cs")
+print(f"Train: {len(cs_dataset['train'])} samples")
+print(f"Validation: {len(cs_dataset['validation'])} samples")
+print(f"Test: {len(cs_dataset['test'])} samples")
+
+# Use cross-view (CV) evaluation protocol
+cv_dataset = load_dataset("omnifall", "cv")
+
+# Use staged-to-wild evaluation protocol (train on staged datasets, test on OOPS)
+staged_to_wild = load_dataset("omnifall", "cs-staged-wild")
+
+# Use individual dataset
+cmdfall = load_dataset("omnifall", "cmdfall")
+
+# Access specific fields from the dataset
+for item in dataset["train"][:5]:
+    print(f"Path: {item['path']}, Label: {item['label']}")
+```
+
+## Experiment Examples
+
+### Cross-Subject Fall Detection
+
+```python
+from datasets import load_dataset
+import torch
+from torch.utils.data import DataLoader
+
+# Load the cross-subject evaluation protocol
+dataset = load_dataset("omnifall", "cs-staged")
+
+# Preprocess and create dataloaders
+def preprocess(examples):
+    # Your preprocessing code here
+    return examples
+
+processed_dataset = dataset.map(preprocess, batched=True)
+train_dataloader = DataLoader(processed_dataset["train"], batch_size=32, shuffle=True)
+val_dataloader = DataLoader(processed_dataset["validation"], batch_size=32)
+test_dataloader = DataLoader(processed_dataset["test"], batch_size=32)
+
+# Train and evaluate your model
+```
+
+### Staged-to-Wild Generalization
+
+```python
+from datasets import load_dataset
+
+# Load the staged-to-wild evaluation protocol
+dataset = load_dataset("omnifall", "cs-staged-wild")
+
+# Train on staged data
+train_data = dataset["train"]
+val_data = dataset["validation"]
+
+# Evaluate on wild data
+wild_test_data = dataset["test"]
+```
+
+## Citation
+
+If you use OmniFall in your research, please cite our paper:
+
+```bibtex
+@inproceedings{omnifall2025,
+  title={OmniFall: A Unified Staged-to-Wild Benchmark for Human Fall Detection},
+  author={},
+  booktitle={},
+  year={2025},
+}
+```
+
+## 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].