bleeding-edge-gameplay-sample

This dataset contains 1024 60 second video clips of Bleeding Edge gameplay (75GB). The data has already been processed into the following format: 300x180 videos sampled at 10 fps.

# Dataset Structure

## Data Files

- testing_dataset_part1.zip & testing_dataset_part2.zip – Contains all 1024 60 second trajectories used for our evaluation.

4 examples from the dataset:
- FB[…].npz – .npz file (described below)
- FB[…].mp4 – 60 seconds .mp4 video of the images from the .npz file. Provided for convenience.

## Data Instances

A typical data entry in the dataset consists of a .npz file that contains uncompressed video frames and corresponding human actions, and accompanying .mp4 file of the same content for easier viewing.

Below is an example from the dataset:

```bash
>>> example_npz["actions"].shape
(610, 16)

>>> example_npz["actions"][:2]
array([[ 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 6., 10., 5., 5.],
[ 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 6., 10., 5., 5.]], dtype=float32)

>>> example_npz["states"].shape
(610, 56)

>>> example_npz["states"][:2]
array([[ 0.5925019 , 0.39726806, 0.478 , 0. , 1. ,
1. , 0.02 , 0.04 , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. ],
[ 0.59672105, 0.39742166, 0.478 , 0. , 1. ,
1. , 0.02 , 0.04 , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. ]], dtype=float32)

>>> example_npz["images"].shape
(610, 3, 180, 300)

>>> example_npz["images"][0]
array([[[185, 183, 183, ..., 121, 120, 121],
[200, 199, 200, ..., 106, 105, 103],
[200, 200, 202, ..., 96, 96, 91],
...,
[159, 160, 149, ..., 128, 127, 136],
[157, 146, 174, ..., 127, 126, 126],
[121, 172, 179, ..., 126, 125, 125]],

[[195, 195, 195, ..., 69, 72, 70],
[210, 211, 212, ..., 74, 75, 74],
[210, 213, 214, ..., 77, 77, 77],
...,
[159, 160, 149, ..., 126, 125, 134],
[157, 146, 174, ..., 127, 125, 125],
[121, 172, 179, ..., 127, 126, 126]],

[[201, 199, 201, ..., 45, 45, 44],
[216, 213, 216, ..., 49, 48, 48],
[216, 214, 218, ..., 52, 50, 50],
...,
[145, 146, 134, ..., 108, 107, 116],
[142, 132, 160, ..., 108, 107, 107],
[107, 158, 164, ..., 108, 107, 107]]], dtype=uint8)
```

An example of an image from the dataset: 

# Dataset Creation

We worked with the game studio Ninja Theory and their game [Bleeding Edge](https://www.bleedingedge.com/) – a 3D, 4v4 multiplayer video game. Recorded data was "replayed" in a special build of the game to render and export it in the form of 300x180 images and corresponding controller actions.

# Source Data

The source data was collected by Ninja Theory, a Xbox Game Studio.

# Data Collection and Processing

The human gameplay data collection and processing was conducted with approval from the Microsoft Institutional Review Board (IRB).

Player data was deidentified prior to use in this research project.

# Limitations

This gameplay data represents gameplay on a single level of the game Bleeding Edge.
This dataset represents gameplay recordings from a variety of skilled and unskilled gameplayers, representing diverse demographic characteristics. Not all possible player characteristics are represented and model performance may therefore vary.
Images in the dataset include watermark and provenance metadata. Do not remove the watermark or provenance metadata.

# License

The model is licensed under the [Microsoft Research License](LICENSE.md).

This work has been funded by Microsoft Research.

# Privacy & Ethics Statement

[Microsoft Privacy Statement](https://go.microsoft.com/fwlink/?LinkId=521839)

# Trademark Notice

**Trademarks** This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow [Microsoft’s Trademark & Brand Guidelines](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general). Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party’s policies.

# Contact Information

For questions please email to [e-mail]

本数据集包含1024段时长为60秒的《嗜血边缘》(Bleeding Edge)游戏游玩录像，总容量75GB。所有数据已预处理为帧率10fps、分辨率300×180的视频格式。

# 数据集结构
## 数据文件
- testing_dataset_part1.zip 与 testing_dataset_part2.zip：包含用于本次评估的全部1024段60秒游玩轨迹数据。

数据集内的4个示例如下：
- FB[…].npz：.npz格式文件（格式说明见下文）
- FB[…].mp4：与该.npz文件对应内容一致的60秒.mp4格式视频，仅为便于查看而提供。

## 数据实例
本数据集的典型数据条目由一个.npz文件与配套的.mp4文件组成：.npz文件存储未压缩的视频帧与对应的人类玩家操作数据，.mp4文件则存储相同内容的视频以供快速预览。

以下为数据集的一个示例：
bash
>>> example_npz["actions"].shape
(610, 16)

>>> example_npz["actions"][:2]
array([[ 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 6., 10., 5., 5.],
[ 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 6., 10., 5., 5.]], dtype=float32)

>>> example_npz["states"].shape
(610, 56)

>>> example_npz["states"][:2]
array([[ 0.5925019 , 0.39726806, 0.478 , 0. , 1. ,
1. , 0.02 , 0.04 , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. ],
[ 0.59672105, 0.39742166, 0.478 , 0. , 1. ,
1. , 0.02 , 0.04 , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. , -1. , -1. , -1. , -1. ,
-1. ]], dtype=float32)

>>> example_npz["images"].shape
(610, 3, 180, 300)

>>> example_npz["images"][0]
array([[[185, 183, 183, ..., 121, 120, 121],
[200, 199, 200, ..., 106, 105, 103],
[200, 200, 202, ..., 96, 96, 91],
...,
[159, 160, 149, ..., 128, 127, 136],
[157, 146, 174, ..., 127, 126, 126],
[121, 172, 179, ..., 126, 125, 125]],

[[195, 195, 195, ..., 69, 72, 70],
[210, 211, 212, ..., 74, 75, 74],
[210, 213, 214, ..., 77, 77, 77],
...,
[159, 160, 149, ..., 126, 125, 134],
[157, 146, 174, ..., 127, 125, 125],
[121, 172, 179, ..., 127, 126, 126]],

[[201, 199, 201, ..., 45, 45, 44],
[216, 213, 216, ..., 49, 48, 48],
[216, 214, 218, ..., 52, 50, 50],
...,
[145, 146, 134, ..., 108, 107, 116],
[142, 132, 160, ..., 108, 107, 107],
[107, 158, 164, ..., 108, 107, 107]]], dtype=uint8)

本数据集的一张示例图像如下：

# 数据集构建
我们与游戏工作室Ninja Theory及其开发的3D 4v4多人对战游戏《嗜血边缘》(Bleeding Edge)[https://www.bleedingedge.com/]开展合作。录制的游戏数据通过该游戏的特殊编译版本进行回放，以渲染并导出为300×180分辨率的图像与对应的控制器操作数据。

# 源数据
源数据由Xbox游戏工作室(Xbox Game Studio)旗下的Ninja Theory收集。

# 数据收集与处理
本次人类游玩数据的收集与处理已通过微软(Microsoft)机构审查委员会(Microsoft Institutional Review Board, IRB)的审批。
在用于本研究项目前，所有玩家数据均已完成匿名化处理。

# 数据集局限性
本数据集仅包含《嗜血边缘》(Bleeding Edge)单张地图的游玩数据。
本数据集涵盖了不同技术水平玩家的游玩录像，具备多样化的人口统计学特征，但未覆盖所有可能的玩家属性，因此模型的实际表现或存在波动。
数据集中的图像包含水印与来源元数据，请勿移除水印及来源元数据。

# 授权协议
本数据集采用[微软研究许可协议(Microsoft Research License)](LICENSE.md)进行授权。
本研究项目由微软研究院(Microsoft Research)资助。

# 隐私与伦理声明
[微软隐私声明(Microsoft Privacy Statement)](https://go.microsoft.com/fwlink/?LinkId=521839)

# 商标声明
**商标** 本项目可能包含项目、产品或服务的商标或徽标。微软(Microsoft)商标与徽标的合法使用需遵守[微软商标与品牌使用指南(Microsoft’s Trademark & Brand Guidelines)](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general)。在本项目的修改版本中使用微软商标或徽标时，不得造成混淆或暗示微软对该修改版本进行赞助。第三方商标与徽标的使用需遵循对应第三方的相关政策。

# 联系方式
如有疑问，请发送邮件至 [e-mail]