File size: 2,739 Bytes
06bd635
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
---
annotations_creators:
- expert-generated
language:
- en
language_creators:
- expert-generated
license: cc-by-4.0
multimodality:
- video
- tabular
pretty_name: Event Human Motion Tracking Dataset
size_categories:
- 100M<n<1B
tags:
- event-based-vision
- neuromorphic
- human-motion-tracking
- computer-vision
- object-tracking
- sensor-bias-optimization
---

# Dataset Card for Event Human Motion Tracking Dataset

## Table of Contents
- [Dataset Description](#dataset-description)
  - [Dataset Summary](#dataset-summary)
  - [Supported Tasks](#supported-tasks)
- [Dataset Structure](#dataset-structure)
  - [Directory Layout](#directory-layout)
  - [File Descriptions](#file-descriptions)
- [Dataset Creation](#dataset-creation)
  - [Sensors & Hardware Setup](#sensors--hardware-setup)
  - [Data Collection & Scenarios](#data-collection--scenarios)
- [Usage & Integration](#usage--integration)
- [Citation](#citation)

---

## Dataset Description

### Dataset Summary
The **Event Human Motion Tracking Dataset** is a dataset designed for fine-grained human motion detection, tracking, and gait/pose analysis. Recorded with a **76° Field of View (FoV)** event camera setup, the dataset explicitly evaluates the performance impact of **Default vs. Optimized sensor bias parameters** on background noise, visual clarity, and object tracking robustness.

The dataset provides synchronized **`.avi` video recordings** for visual inspection, algorithmic benchmarking, and spatial ground-truth verification across both bias configurations.

### Supported Tasks
* **Human Motion Tracking:** Bounding box detection and temporal association across dynamic motion recordings.
* **Sensor Bias Benchmarking:** Comparative evaluation of default vs. custom-optimized sensor bias settings ($76^\circ$ FoV).
* **Neuromorphic Motion Analysis:** Feature extraction, velocity estimation, and performance assessment under varied bias regimes.

---

## Dataset Structure

### Directory Layout
The repository contains `.avi` video recordings separated by bias configuration, alongside compressed archives for annotations and parameter settings:

```text
Event-Human-Motion-Tracking-Dataset/
├── Default-Bias-Data-76-Degree-FoV/     # Video recordings captured with Default Bias (.avi)
│   ├── recording_01.avi
│   └── recording_02.avi
├── Optimized-Bias-Data-76-Degree-FoV/   # Video recordings captured with Optimized Bias (.avi)
│   ├── recording_01.avi
│   └── recording_02.avi
├── Bias-Settings-Real-Time.zip          # Real-time sensor bias configuration logs and register settings
├── Mixed-Bias-Data-Annotations.zip      # Ground truth bounding box annotations and track IDs
└── README.md