PFED5+
PFED5+是用于帕金森病面部表情质量评估的扩展视频数据集，含2,811个视频片段及专家注释的运动描述文本，支持生成式可解释性分析。
基本信息
资源简介
PFED5+
【数据集背景】
本数据集源站标题为《shuchaoduan/TraMP-LLaMA》。源站首次发布：2026-06-19。
【数据内容】
源站原始描述：TraMP-LLaMA: Generative Interpretability with Decoupled Instruction Tuning for Facial Expression Quality Assessment # TraMP-LLaMA TraMP-LLaMA: Generative Interpretability with Decoupled Instruction Tuning for Facial Expression Quality Assessment arXiv version We propose TraMP-LLaMA, a unified multimodal framework that jointly predicts severity scores and generates structured textual reports from facial motion cues. To support this task, we further extend the PFED5 dataset with expert-guided text…
【适用方向】
多模态数据适合跨模态对齐、融合建模与联合表征学习方向的研究。
【获取与许可】
源站页面：https://github.com/shuchaoduan/TraMP-LLaMA
使用许可：MIT。
【使用注意】
· 源站最后修改于 2026-06-26，引用时请注意版本时效。
下载信息
注册下载
需要注册 Kaggle 账号并登录后下载，适合需要跟踪下载记录和使用 API 的用户。
暂未开放公开下载
无需注册即可直接获取公开样本或文档，适合快速预览和评估数据集质量。
免登录有偿下载
公开数据集受托下载与技术交付服务。
提供高速下载与技术交付服务（收技术服务费，非数据销售）
暂未开放当前数据集主要来源为 Kaggle 公开托管，完整影像包建议通过原始链接或 Kaggle API 获取。
使用方式
数据集获取
git clone https://github.com/shuchaoduan/TraMP-LLaMA.git
curl -L -o repo.zip https://github.com/shuchaoduan/TraMP-LLaMA/archive/refs/heads/main.zip
unzip repo.zip
源站 README 摘录（使用方式）
TraMP-LLaMA
TraMP-LLaMA: Generative Interpretability with Decoupled Instruction Tuning for Facial Expression Quality Assessment
arXiv version
We propose TraMP-LLaMA, a unified multimodal framework that jointly predicts severity scores and generates structured textual reports from facial motion cues. To support this task, we further extend the PFED5 dataset with expert-guided textual motion descriptions and construct PFED5+.
PFED5+ dataset
1: Please request the access to video frames and MDS-UPDRS labels through this link.
2: The motion description labels are provided here.
Citations
If you find our work useful in your research, please consider giving it a star ⭐ and citing our paper in your work:
@misc{trampllama,
title={TraMP-LLaMA: Generative Interpretability with Decoupled Instruction Tuning for Facial Expression Quality Assessment},
author={Shuchao Duan and Alan Whone and Hossein Rahmani and Jun Liu and Majid Mirmehdi},
year={2026},
eprint={2606.26942},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
@INPROCEEDINGS{tramp-former,
title={Trajectory-guided Motion Perception for Facial Expression Quality Assessment in Neurological Disorders},
author={Shuchao Duan and Amirhossein Dadashzadeh and Alan Whone and Majid Mirmehdi},
booktitle={2025 IEEE 19th International Conference on Automatic Face and Gesture Recognition (FG)},
year={2025},
doi={10.1109/FG61629.2025.11099263}
}
@misc{duan2023qafenet,
title={QAFE-Net: Quality Assessment of Facial Expressions with Landmark Heatmaps},
author={Shuchao Duan and Amirhossein Dadashzadeh and Alan Whone and Majid Mirmehdi},
year={2023},
eprint={2312.00856},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
Acknowledgement
This codebase builds upon VideoLLaMA3. We gratefully acknowledge their contributions to the open-source community.
数据加载示例（表格/文本类）
import pandas as pd, glob, os
files = (glob.glob(os.path.join(path, "**", "*.csv"), recursive=True)
+ glob.glob(os.path.join(path, "**", "*.tsv"), recursive=True)
+ glob.glob(os.path.join(path, "**", "*.xlsx"), recursive=True))
print("数据文件:", files)
df = pd.read_csv(files[0])
print(df.shape); print(df.columns.tolist()); print(df.head(3))
数据缺失？
依托陕西、四川两大基地，我们与超过十家三甲医院建立直接合作关系，覆盖合规授权、采集治理、专业标注、数据交付的全流程，为AI医疗团队提供即拿即用的高质量临床数据。
⚡️ 需要数据支持或标注服务？立即联系我们获取专业方案。
