catandsoda/KABR
收藏Hugging Face2026-04-14 更新2026-04-26 收录
下载链接：
https://hf-mirror.com/datasets/catandsoda/KABR
下载链接
链接失效反馈官方服务：
资源简介：
---
license: cc0-1.0
task_categories:
- video-classification
tags:
- zebra
- giraffe
- plains zebra
- Grevy's zebra
- video
- animal behavior
- behavior recognition
- annotation
- annotated video
- conservation
- drone
- UAV
- imbalanced
- Kenya
- Mpala Research Centre
pretty_name: >-
KABR: In-Situ Dataset for Kenyan Animal Behavior Recognition from Drone
Videos
description: "Initial KABR project release, contains drone video clips (mini-scenes) of giraffes, plains zebras, and Grevy's zebras with behavior labels from a subset of videos collected at the Mpala Research Centre in January 2023."
size_categories:
- 1M<n<10M
---
# Dataset Card for KABR: In-Situ Dataset for Kenyan Animal Behavior Recognition from Drone Videos
## Dataset Description
- **Homepage:** [KABR Mini-Scene Site](https://kabrdata.xyz/)
- **Project Page:** [KABR Site](https://imageomics.github.io/KABR/)
- **Repository:** https://github.com/Imageomics/kabr-tools
- **Paper:** https://openaccess.thecvf.com/content/WACV2024W/CV4Smalls/papers/Kholiavchenko_KABR_In-Situ_Dataset_for_Kenyan_Animal_Behavior_Recognition_From_Drone_WACVW_2024_paper.pdf
### Dataset Summary
We present a novel high-quality dataset for animal behavior recognition from drone videos.
The dataset is focused on Kenyan wildlife and contains behaviors of giraffes, plains zebras, and Grevy's zebras.
The dataset consists of more than 10 hours of annotated videos, and it includes eight different classes, encompassing seven types of animal behavior and an additional category for occluded instances.
In the annotation process for this dataset, a team of 10 people was involved, with an expert zoologist overseeing the process.
Each behavior was labeled based on its distinctive features, using a standardized set of criteria to ensure consistency and accuracy across the annotations.
The dataset was collected using drones that flew over the animals in the [Mpala Research Centre](https://mpala.org/) in Kenya, providing high-quality video footage of the animal's natural behaviors.
The drone footage is captured at a resolution of 5472 x 3078 pixels, and the videos were recorded at a frame rate of 29.97 frames per second.
<!--This dataset card aims to be a base template for new datasets. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/datasetcard_template.md?plain=1).-->
### Supported Tasks and Leaderboards
The results of our evaluation using I3D, SlowFast, and X3D architectures are given in the table below. For each one, the model was trained for 120 epochs with batch size of 5. For more information on these results, see our [paper](coming soon).
| Method | All | Giraffes | Plains Zebras | Grevy’s Zebras |
| ---- | ---- | ---- | ---- | ---- |
| I3D (16x5) | 53.41 | 61.82 | 58.75 | 46.73 |
| SlowFast (16x5, 4x5) | 52.92 | 61.15 | 60.60 | 47.42 |
| X3D (16x5) | 61.9 | 65.1 | 63.11 | 51.16 |
### Languages
English
## Dataset Structure
Under `KABR/dataset/image/`, the data has been archived into `.zip` files, which are split into 2GB files. These must be recombined and extracted.
After cloning and navigating into the repository, you can use the following commands to do the reconstruction:
```bash
cd KABR/dataset/image/
cat giraffes_part_* > giraffes.zip
md5sum giraffes.zip # Compare this to what's shown with `cat giraffes_md5.txt`
unzip giraffes.zip
rm -rf giraffes_part_*
# Similarly for `zebras_grevys_part_*` and `zebras_plains_part_*`
```
Alternatively, there is a download script, `download.py`, which allows a download of the entire dataset in its established format without requiring one to clone the repository (cloning requires _at least_ double the size of the dataset to store). To proceed with this approach, download `download.py` to the system where you want to access the data.
Then, in the same directory as the script, run the following to begin the download:
```
pip install requests
python download.py
```
This script then downloads all the files present in the repository (without making a clone of the `.git` directory, etc.), concatenates the part files to their ZIP archives, verifies the MD5 checksums, extracts, and cleans up so that the folder structure, as described below, is present.
Note that it will require approximately 116GB of free space to complete this process, though the final dataset will only take about 61GB of disk space (the script removes the extra files after checking the download was successful).
The KABR dataset follows the Charades format:
```
KABR
/dataset
/image
/video_1
/image_1.jpg
/image_2.jpg
...
/image_n.jpg
/video_2
/image_1.jpg
/image_2.jpg
...
/image_n.jpg
...
/video_n
/image_1.jpg
/image_2.jpg
/image_3.jpg
...
/image_n.jpg
/annotation
/classes.json
/train.csv
/val.csv
```
The dataset can be directly loaded and processed by the [SlowFast](https://github.com/facebookresearch/SlowFast) framework.
**Informational Files**
* `KABR/configs`: examples of SlowFast framework configs.
* `KABR/annotation/distribution.xlsx`: distribution of classes for all videos.
**Scripts:**
* `image2video.py`: Encode image sequences into the original video.
* For example, `[image/G0067.1, image/G0067.2, ..., image/G0067.24]` will be encoded into `video/G0067.mp4`.
* `image2visual.py`: Encode image sequences into the original video with corresponding annotations.
* For example, `[image/G0067.1, image/G0067.2, ..., image/G0067.24]` will be encoded into `visual/G0067.mp4`.
### Data Instances
**Naming:** Within the image folder, the `video_n` folders are named as follows (X indicates a number):
* G0XXX.X - Giraffes
* ZP0XXX.X - Plains Zebras
* ZG0XXX.X - Grevy's Zebras
* Within each of these folders the images are simply `X.jpg`.
**Note:** The dataset consists of a total of 1,139,893 frames captured from drone videos. There are 488,638 frames of Grevy's zebras, 492,507 frames of plains zebras, and 158,748 frames of giraffes.
### Data Fields
There are 14,764 unique behavioral sequences in the dataset. These consist of eight distinct behaviors:
- Walk
- Trot
- Run: animal is moving at a cantor or gallop
- Graze: animal is eating grass or other vegetation
- Browse: animal is eating trees or bushes
- Head Up: animal is looking around or observe surroundings
- Auto-Groom: animal is grooming itself (licking, scratching, or rubbing)
- Occluded: animal is not fully visible
### Data Splits
Training and validation sets are indicated by their respective CSV files (`train.csv` and `val.csv`), located within the `annotation` folder.
## Dataset Creation
### Curation Rationale
We present a novel high-quality dataset for animal behavior recognition from drone videos.
The dataset is focused on Kenyan wildlife and contains behaviors of giraffes, plains zebras, and Grevy's zebras.
The dataset consists of more than 10 hours of annotated videos, and it includes eight different classes, encompassing seven types of animal behavior and an additional category for occluded instances.
In the annotation process for this dataset, a team of 10 people was involved, with an expert zoologist overseeing the process.
Each behavior was labeled based on its distinctive features, using a standardized set of criteria to ensure consistency and accuracy across the annotations.
The dataset was collected using drones that flew over the animals in the [Mpala Research Centre](https://mpala.org/) in Kenya, providing high-quality video footage of the animal's natural behaviors.
We believe that this dataset will be a valuable resource for researchers working on animal behavior recognition, as it provides a diverse and high-quality set of annotated videos that can be used for evaluating deep learning models.
Additionally, the dataset can be used to study the behavior patterns of Kenyan animals and can help to inform conservation efforts and wildlife management strategies.
<!-- [To be added:] -->
We provide a detailed description of the dataset and its annotation process, along with some initial experiments on the dataset using conventional deep learning models.
The results demonstrate the effectiveness of the dataset for animal behavior recognition and highlight the potential for further research in this area.
### Source Data
#### Initial Data Collection and Normalization
Data was collected from 6 January 2023 through 21 January 2023 at the [Mpala Research Centre](https://mpala.org/) in Kenya under a Nacosti research license. We used DJI Mavic 2S drones equipped with cameras to record 5.4K resolution videos (5472 x 3078 pixels) from varying altitudes and distances of 10 to 50 meters from the animals (distance was determined by circumstances and safety regulations).
Mini-scenes were extracted from these videos to reduce the impact of drone movement and facilitate human annotation. Animals were detected in frame using YOLOv8, then the SORT tracking algorithm was applied to follow their movement. A 400 by 300 pixel window, centered on the animal, was then extracted; this is the mini-scene.
<!--
#### Who are the source language producers?
[More Information Needed]
-->
### Annotations
#### Annotation process
In the annotation process for this dataset, a team of 10 people was involved, with an expert zoologist overseeing the process.
Each behavior was labeled based on its distinctive features, using a standardized set of criteria to ensure consistency and accuracy across the annotations.
<!--
#### Who are the annotators?
[More Information Needed]
-->
### Personal and Sensitive Information
Though there are endangered species included in this data, exact locations are not provided and their safety is assured by their location within the preserve.
## Considerations for Using the Data
<!--
### Social Impact of Dataset
[More Information Needed]
### Discussion of Biases
[More Information Needed]
-->
### Other Known Limitations
This data exhibits a long-tailed distribution due to the natural variation in frequency of the observed behaviors.
## Additional Information
### Authors
* Maksim Kholiavchenko (Rensselaer Polytechnic Institute) - ORCID: 0000-0001-6757-1957
* Jenna Kline (The Ohio State University) - ORCID: 0009-0006-7301-5774
* Michelle Ramirez (The Ohio State University)
* Sam Stevens (The Ohio State University)
* Alec Sheets (The Ohio State University) - ORCID: 0000-0002-3737-1484
* Reshma Ramesh Babu (The Ohio State University) - ORCID: 0000-0002-2517-5347
* Namrata Banerji (The Ohio State University) - ORCID: 0000-0001-6813-0010
* Elizabeth Campolongo (Imageomics Institute, The Ohio State University) - ORCID: 0000-0003-0846-2413
* Matthew Thompson (Imageomics Institute, The Ohio State University) - ORCID: 0000-0003-0583-8585
* Nina Van Tiel (Eidgenössische Technische Hochschule Zürich) - ORCID: 0000-0001-6393-5629
* Jackson Miliko (Mpala Research Centre)
* Eduardo Bessa (Universidade de Brasília) - ORCID: 0000-0003-0606-5860
* Tanya Berger-Wolf (The Ohio State University) - ORCID: 0000-0001-7610-1412
* Daniel Rubenstein (Princeton University) - ORCID: 0000-0001-9049-5219
* Charles Stewart (Rensselaer Polytechnic Institute)
### Licensing Information
This dataset is dedicated to the public domain for the benefit of scientific pursuits. We ask that you cite the dataset and journal paper using the below citations if you make use of it in your research.
### Citation Information
#### Dataset
```
@misc{KABR_Data,
author = {Kholiavchenko, Maksim and Kline, Jenna and Ramirez, Michelle and Stevens, Sam and Sheets, Alec and Babu, Reshma and Banerji, Namrata and Campolongo, Elizabeth and Thompson, Matthew and Van Tiel, Nina and Miliko, Jackson and Bessa, Eduardo and Duporge, Isla and Berger-Wolf, Tanya and Rubenstein, Daniel and Stewart, Charles},
title = {KABR: In-Situ Dataset for Kenyan Animal Behavior Recognition from Drone Videos},
year = {2023},
url = {https://huggingface.co/datasets/imageomics/KABR},
doi = {10.57967/hf/1010},
publisher = {Hugging Face}
}
```
#### Paper
```
@inproceedings{kholiavchenko2024kabr,
title={KABR: In-Situ Dataset for Kenyan Animal Behavior Recognition from Drone Videos},
author={Kholiavchenko, Maksim and Kline, Jenna and Ramirez, Michelle and Stevens, Sam and Sheets, Alec and Babu, Reshma and Banerji, Namrata and Campolongo, Elizabeth and Thompson, Matthew and Van Tiel, Nina and Miliko, Jackson and Bessa, Eduardo and Duporge, Isla and Berger-Wolf, Tanya and Rubenstein, Daniel and Stewart, Charles},
booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
pages={31-40},
year={2024}
}
```
### Contributions
This work was supported by the [Imageomics Institute](https://imageomics.org), which is funded by the US National Science Foundation's Harnessing the Data Revolution (HDR) program under [Award #2118240](https://www.nsf.gov/awardsearch/showAward?AWD_ID=2118240) (Imageomics: A New Frontier of Biological Information Powered by Knowledge-Guided Machine Learning). Additional support was also provided by the [AI Institute for Intelligent Cyberinfrastructure with Computational Learning in the Environment (ICICLE)](https://icicle.osu.edu/), which is funded by the US National Science Foundation under [Award #2112606](https://www.nsf.gov/awardsearch/showAward?AWD_ID=2112606). Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
The data was gathered at the [Mpala Research Centre](https://mpala.org/) in Kenya, in accordance with Research License No. NACOSTI/P/22/18214. The data collection protocol adhered strictly to the guidelines set forth by the Institutional Animal Care and Use Committee under permission No. IACUC 1835F.
许可证：CC0 1.0协议
任务类别：
- 视频分类（video-classification）
标签：
- 斑马（zebra）
- 长颈鹿（giraffe）
- 平原斑马（plains zebra）
- 格氏斑马（Grevy's zebra）
- 视频（video）
- 动物行为（animal behavior）
- 行为识别（behavior recognition）
- 标注（annotation）
- 标注视频（annotated video）
- 保护（conservation）
- 无人机（drone）
- 无人机（UAV）
- 样本不平衡（imbalanced）
- 肯尼亚（Kenya）
- 姆帕拉研究中心（Mpala Research Centre）
数据集展示名称：KABR：基于无人机视频的肯尼亚动物行为识别原位数据集（KABR: In-Situ Dataset for Kenyan Animal Behavior Recognition from Drone Videos）
数据集描述：本内容为KABR项目的初始发布版本，包含2023年1月于姆帕拉研究中心采集的部分视频中提取的长颈鹿、平原斑马、格氏斑马的无人机视频片段（迷你场景），并附带行为标注。
样本规模区间：
- 100万<n<1000万
# KABR数据集卡片：基于无人机视频的肯尼亚动物行为识别原位数据集
## 数据集描述
- "**主页**"：[KABR迷你场景站点](https://kabrdata.xyz/)
- "**项目页面**"：[KABR站点](https://imageomics.github.io/KABR/)
- "**代码仓库**"：https://github.com/Imageomics/kabr-tools
- "**论文**"：https://openaccess.thecvf.com/content/WACV2024W/CV4Smalls/papers/Kholiavchenko_KABR_In-Situ_Dataset_for_Kenyan_Animal_Behavior_Recognition_From_Drone_WACVW_2024_paper.pdf
### 数据集摘要
本数据集面向无人机视频中的动物行为识别任务，是一款全新的高质量数据集。该数据集聚焦肯尼亚野生动物，涵盖长颈鹿、平原斑马与格氏斑马的行为数据。数据集包含超过10小时的标注视频，共包含8个类别：7类动物行为与1类遮挡实例。本数据集的标注工作由10人团队完成，并由一名动物学专家全程监督。所有行为标注均基于其独特特征，采用标准化的标注准则以确保全量标注的一致性与准确性。数据集采集自肯尼亚姆帕拉研究中心，通过无人机飞越动物栖息地获取，保留了动物自然行为的高质量视频素材。该无人机视频的分辨率为5472×3078像素，帧率为29.97帧每秒。
### 支持任务与基准榜单
我们使用I3D、SlowFast与X3D架构开展评估，评估结果如下表所示。所有模型均以批次大小5训练120个epoch。更多评估细节详见我们的[论文（即将上线）](coming soon)。
| 方法 | 全部类别 | 长颈鹿 | 平原斑马 | 格氏斑马 |
| ---- | ---- | ---- | ---- | ---- |
| I3D (16x5) | 53.41 | 61.82 | 58.75 | 46.73 |
| SlowFast (16x5, 4x5) | 52.92 | 61.15 | 60.60 | 47.42 |
| X3D (16x5) | 61.9 | 65.1 | 63.11 | 51.16 |
### 语言
英语
## 数据集结构
数据归档于`KABR/dataset/image/`路径下，被拆分为多个2GB的分卷zip文件，需先合并再解压。克隆仓库并进入对应目录后，可使用以下命令完成合并：
bash
cd KABR/dataset/image/
cat giraffes_part_* > giraffes.zip
md5sum giraffes.zip # 将输出结果与`cat giraffes_md5.txt`中的值进行比对
unzip giraffes.zip
rm -rf giraffes_part_*
# 对`zebras_grevys_part_*`与`zebras_plains_part_*`执行相同操作
此外，项目提供了下载脚本`download.py`，无需克隆仓库即可直接下载完整数据集（克隆仓库所需的存储空间至少为数据集大小的两倍）。使用该方法的步骤如下：
1. 将`download.py`下载至需要使用数据集的系统中
2. 在脚本所在目录执行以下命令开始下载：
pip install requests
python download.py
该脚本会下载仓库中的所有文件（无需克隆`.git`目录等额外内容），自动合并分卷文件为zip归档，验证MD5校验和，解压并清理临时文件，最终得到如下所述的文件夹结构。
该过程需约116GB的可用存储空间，最终数据集仅占用约61GB磁盘空间（脚本会在验证下载成功后删除临时文件）。
KABR数据集遵循Charades格式，目录结构如下：
KABR
/dataset
/image
/video_1
/image_1.jpg
/image_2.jpg
...
/image_n.jpg
/video_2
/image_1.jpg
/image_2.jpg
...
/image_n.jpg
...
/video_n
/image_1.jpg
/image_2.jpg
/image_3.jpg
...
/image_n.jpg
/annotation
/classes.json
/train.csv
/val.csv
该数据集可直接通过[SlowFast](https://github.com/facebookresearch/SlowFast)框架加载与处理。
"**信息文件**"
* `KABR/configs`：SlowFast框架的配置示例文件
* `KABR/annotation/distribution.xlsx`：所有视频的类别分布统计
"**脚本**"
* `image2video.py`：将图像序列编码为原始视频。例如，将`[image/G0067.1, image/G0067.2, ..., image/G0067.24]`编码为`video/G0067.mp4`。
* `image2visual.py`：将图像序列编码为带有对应标注的可视化视频。例如，将`[image/G0067.1, image/G0067.2, ..., image/G0067.24]`编码为`visual/G0067.mp4`。
### 数据实例
"**命名规则**"：在图像文件夹中，`video_n`文件夹的命名格式如下（X为数字）：
* G0XXX.X - 长颈鹿
* ZP0XXX.X - 平原斑马
* ZG0XXX.X - 格氏斑马
每个文件夹内的图像均以`X.jpg`命名。
"**备注**"：本数据集共包含1,139,893帧无人机视频素材，其中格氏斑马488,638帧，平原斑马492,507帧，长颈鹿158,748帧。
### 数据字段
本数据集共包含14,764个独特的行为序列，涵盖8个类别：
- 行走（Walk）
- 小跑（Trot）
- 奔跑（Run：动物以慢跑或疾驰的方式移动）
- 啃食草本植物（Graze：动物食用青草或其他植被）
- 啃食木本植物（Browse：动物食用树木或灌木）
- 抬头观察（Head Up：动物环顾四周或观察周边环境）
- 自我理毛（Auto-Groom：动物通过舔舐、抓挠或摩擦进行自我梳理）
- 遮挡（Occluded：动物未完全可见）
### 数据划分
训练集与验证集由`annotation`文件夹下的`train.csv`与`val.csv`分别指定。
## 数据集创建
### 策划依据
我们推出这款面向无人机视频动物行为识别的全新高质量数据集，聚焦肯尼亚野生动物，涵盖长颈鹿、平原斑马与格氏斑马的行为数据。数据集包含超过10小时的标注视频，共8个类别：7类动物行为与1类遮挡实例。标注工作由10人团队完成，由动物学专家全程监督，所有行为标注均基于其独特特征，采用标准化准则以确保标注的一致性与准确性。数据集采集自肯尼亚姆帕拉研究中心，通过无人机飞越动物栖息地获取，保留了动物自然行为的高质量视频素材。我们认为，该数据集将为从事动物行为识别研究的人员提供宝贵的资源，其包含的多样化高质量标注视频可用于评估深度学习模型。此外，该数据集还可用于研究肯尼亚野生动物的行为模式，为保护工作与野生动物管理策略提供参考依据。
我们提供了数据集及其标注流程的详细描述，并基于该数据集开展了使用传统深度学习模型的初步实验。实验结果证明了该数据集在动物行为识别任务中的有效性，同时也凸显了该领域进一步研究的潜力。
### 源数据
#### 初始数据采集与归一化
数据采集于2023年1月6日至1月21日的肯尼亚姆帕拉研究中心，持有肯尼亚国家科学、技术与创新委员会（Nacosti）研究许可。研究团队使用搭载相机的DJI Mavic 2S无人机，以5.4K分辨率（5472×3078像素）录制视频，飞行高度与动物的距离为10至50米（距离由实际场景与安全规范决定）。
研究团队从原始视频中提取迷你场景，以降低无人机运动的影响并便于人工标注：首先使用YOLOv8检测帧中的动物，再通过SORT跟踪算法追踪其运动轨迹，最终以动物为中心提取400×300像素的窗口，即为本数据集的迷你场景。
### 标注
#### 标注流程
本数据集的标注工作由10人团队完成，并由一名动物学专家全程监督。所有行为标注均基于其独特特征，采用标准化的标注准则以确保全量标注的一致性与准确性。
### 个人与敏感信息
尽管数据中包含濒危物种，但未提供精确位置，且所有数据均采集于保护区内，确保了动物的安全。
## 数据使用注意事项
### 其他已知局限性
由于观测到的行为频率存在自然差异，该数据呈现长尾分布。
## 补充信息
### 作者
* 马克西姆·霍利亚夫琴科（伦斯勒理工学院） - ORCID：0000-0001-6757-1957
* 珍娜·克莱因（俄亥俄州立大学） - ORCID：0009-0006-7301-5774
* 米歇尔·拉米雷斯（俄亥俄州立大学）
* 萨姆·史蒂文斯（俄亥俄州立大学）
* 亚历克·希茨（俄亥俄州立大学） - ORCID：0000-0002-3737-1484
* 雷斯玛·拉梅什·巴布（俄亥俄州立大学） - ORCID：0000-0002-2517-5347
* 纳姆拉塔·班纳吉（俄亥俄州立大学） - ORCID：0000-0001-6813-0010
* 伊丽莎白·坎波隆戈（图像组学研究所，俄亥俄州立大学） - ORCID：0000-0003-0846-2413
* 马修·汤普森（图像组学研究所，俄亥俄州立大学） - ORCID：0000-0003-0583-8585
* 妮娜·范蒂尔（苏黎世联邦理工学院） - ORCID：0000-0001-6393-5629
* 杰克逊·米利科（姆帕拉研究中心）
* 爱德华多·贝萨（巴西利亚大学） - ORCID：0000-0003-0606-5860
* 坦尼娅·伯杰-沃尔夫（俄亥俄州立大学） - ORCID：0000-0001-7610-1412
* 丹尼尔·鲁宾斯坦（普林斯顿大学） - ORCID：0000-0001-9049-5219
* 查尔斯·斯图尔特（伦斯勒理工学院）
### 许可信息
本数据集已奉献至公共领域，以服务于科学研究。若在研究中使用该数据集，请引用下方的数据集与期刊论文信息。
### 引用信息
#### 数据集
@misc{KABR_Data,
author = {Kholiavchenko, Maksim and Kline, Jenna and Ramirez, Michelle and Stevens, Sam and Sheets, Alec and Babu, Reshma and Banerji, Namrata and Campolongo, Elizabeth and Thompson, Matthew and Van Tiel, Nina and Miliko, Jackson and Bessa, Eduardo and Duporge, Isla and Berger-Wolf, Tanya and Rubenstein, Daniel and Stewart, Charles},
title = {KABR: In-Situ Dataset for Kenyan Animal Behavior Recognition from Drone Videos},
year = {2023},
url = {https://huggingface.co/datasets/imageomics/KABR},
doi = {10.57967/hf/1010},
publisher = {Hugging Face}
}
#### 期刊论文
@inproceedings{kholiavchenko2024kabr,
title={KABR: In-Situ Dataset for Kenyan Animal Behavior Recognition from Drone Videos},
author={Kholiavchenko, Maksim and Kline, Jenna and Ramirez, Michelle and Stevens, Sam and Sheets, Alec and Babu, Reshma and Banerji, Namrata and Campolongo, Elizabeth and Thompson, Matthew and Van Tiel, Nina and Miliko, Jackson and Bessa, Eduardo and Duporge, Isla and Berger-Wolf, Tanya and Rubenstein, Daniel and Stewart, Charles},
booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
pages={31-40},
year={2024}
}
### 资助信息
本工作由图像组学研究所（Imageomics Institute）资助，该研究所由美国国家科学基金会（NSF）的“驾驭数据革命”（Harnessing the Data Revolution, HDR）计划下的奖项#2118240资助（项目名称：Imageomics：以知识引导机器学习推动生物信息学新前沿）。额外资助来自智能网络基础设施与计算学习环境AI研究所（AI Institute for Intelligent Cyberinfrastructure with Computational Learning in the Environment, ICICLE），该研究所由美国国家科学基金会通过奖项#2112606资助。本材料中的任何观点、发现、结论或建议均为作者本人的观点，不一定反映国家科学基金会的立场。
数据采集于肯尼亚姆帕拉研究中心，遵循研究许可编号NACOSTI/P/22/18214，且严格遵循机构动物护理与使用委员会的指南，许可编号为IACUC 1835F。
提供机构：
catandsoda
