DescriptionThis dataset contains a collection of synthetic hand images and corresponding annotation files, designed for machine learning tasks such as keypoint detection, landmark localization, and acupoint-related research. The dataset includes both original-resolution and resized RGB images, together with structured annotations in CSV and JSON formats.The MetaAcuPoint dataset was originally introduced in the following paper:Journal Article How to Cite: If you use this dataset in your research, please cite the following:@article{guruge_metaacupoint_2025,  author  = {Guruge, K. and Padmanabha, P. and Herath, H. M. K. K. M. B. and Madusanka, N. and Park, H.-J. and Na, C.-S. and Yi, M. and Lee, B.},  title   = {MetaAcuPoint: MetaHuman-Generated Synthetic Data for Hand Acupoint Localization},  journal = {Healthcare},  year    = {2025},  volume  = {13},  number  = {23},  pages   = {3093},  doi     = {10.3390/healthcare13233093},  url     = {https://doi.org/10.3390/healthcare13233093}}@dataset{guruge2025metaacupoint,  author       = {Guruge, P. K. and Padmanabha, P. and Herath, H. M. K. K. M. B. and Vithanage, N. M. and Park, H.-J. and Na, C. and Yi, M. and Lee, B.},  title        = {MetaAcuPoint: MetaHuman-Generated Synthetic Forearm Data},  year         = {2025},  publisher    = {Zenodo},  version      = {v1.0},  howpublished = {In \textit{MetaAcuPoint: MetaHuman-Generated Synthetic Data for Hand Acupoint Localization}},  doi          = {10.5281/zenodo.17713204},  url          = {https://doi.org/10.5281/zenodo.17713204}}orGuruge, K., Padmanabha, P., Herath, H. M. K. K. M. B., Madusanka, N., Park, H.-J., Na, C.-S., Yi, M., & Lee, B. (2025). MetaAcuPoint: MetaHuman-Generated Synthetic Data for Hand Acupoint Localization. Healthcare, 13(23), 3093. https://doi.org/10.3390/healthcare13233093.   Guruge, P. K., Padmanabha, P., Herath, H. M. K. K. M. B., Vithanage, N. M., Park, H.-J., Na, C., Yi, M., & Lee, B. (2025). MetaAcuPoint: MetaHuman-Generated Synthetic Forearm Data [Data set]. In MetaAcuPoint: MetaHuman-Generated Synthetic Data for Hand Acupoint Localization (v1.0). Zenodo. https://doi.org/10.5281/zenodo.17713204Dataset ContentsThe dataset consists of the following components:1. original_RGBA folder containing the 900 full-resolution RGB images generated under controlled lighting and consistent top-down imaging geometry.These images serve as the baseline dataset for training or benchmarking high-precision localization models.2. resized_RGBA folder containing the 900 resized versions of the original images. Resizing was performed to match input constraints of deep learning frameworks or to standardize spatial dimensions for downstream analysis. All resized images maintain the original aspect ratio and image quality suitable for annotation.3. annotation_Original_RGB.csvA comma-separated annotation file providing pixel-level labels for keypoints or region-of-interest coordinates corresponding to each image in the original_RGB folder.Each row contains:Image filenameX, Y coordinate values for the designated keypoints4. annotation_resized_RGB.jsonA JSON-formatted annotation file corresponding to the resized_RGB images. It follows a COCO-style schema, containing:Image metadata (filename, height, width)Keypoint annotationsSegmentation or bounding box fieldsCategory definitionsThis format is directly compatible with popular computer vision libraries such as MMPose and COCO API.5. avatar_description.xlsxAn Excel file containing descriptive information about the metahuman avatars included in the dataset. This includes demographic information, avatar attributes, and complementary metadata.Dataset Folder Structure (Unzip dataset_MetaAcuPoint.zip)dataset_MetaAcuPoint/│├── original_RGB/│   ├── img_0001.png│   ├── img_0002.png│   └── ...│├── resized_RGB/│   ├── img_0001.png│   ├── img_0002.png│   └── ...│├── annotation_Original_RGB.csv├── annotation_resized_RGB.json├── avatar_description.xlsx└── README.mdImage Naming ConventionEach image in the dataset follows a structured naming format: [AvatarID]arm[HandSide][FrameNumber].png | [XXX]arm[Y][Z].pngNote:AvatarID (XXX)A three-digit identifier ranging from 101 to 130Example: 101, 115, 130HandSide (X)Indicates which hand is shown    1 → Right hand    2 → Left handFrameNumber (Z)Frame index from 1 to 15, corresponding to the sequence extracted from the MetaHuman animation.

DescriptionThis dataset contains a collection of synthetic hand images and corresponding annotation files, designed for machine learning tasks such as keypoint detection, landmark localization, and acupoint-related research. The dataset includes both original-resolution and resized RGB images, together with structured annotations in CSV and JSON formats.The MetaAcuPoint dataset was originally introduced in the following paper:Journal Article How to Cite: If you use this dataset in your research, please cite the following:@article{guruge_metaacupoint_2025,  author  = {Guruge, K. and Padmanabha, P. and Herath, H. M. K. K. M. B. and Madusanka, N. and Park, H.-J. and Na, C.-S. and Yi, M. and Lee, B.},  title   = {MetaAcuPoint: MetaHuman-Generated Synthetic Data for Hand Acupoint Localization},  journal = {Healthcare},  year    = {2025},  volume  = {13},  number  = {23},  pages   = {3093},  doi     = {10.3390/healthcare13233093},  url     = {https://doi.org/10.3390/healthcare13233093}}@dataset{guruge2025metaacupoint,  author       = {Guruge, P. K. and Padmanabha, P. and Herath, H. M. K. K. M. B. and Vithanage, N. M. and Park, H.-J. and Na, C. and Yi, M. and Lee, B.},  title        = {MetaAcuPoint: MetaHuman-Generated Synthetic Forearm Data},  year         = {2025},  publisher    = {Zenodo},  version      = {v1.0},  howpublished = {In \textit{MetaAcuPoint: MetaHuman-Generated Synthetic Data for Hand Acupoint Localization}},  doi          = {10.5281/zenodo.17713204},  url          = {https://doi.org/10.5281/zenodo.17713204}}orGuruge, K., Padmanabha, P., Herath, H. M. K. K. M. B., Madusanka, N., Park, H.-J., Na, C.-S., Yi, M., & Lee, B. (2025). MetaAcuPoint: MetaHuman-Generated Synthetic Data for Hand Acupoint Localization. Healthcare, 13(23), 3093. https://doi.org/10.3390/healthcare13233093.   Guruge, P. K., Padmanabha, P., Herath, H. M. K. K. M. B., Vithanage, N. M., Park, H.-J., Na, C., Yi, M., & Lee, B. (2025). MetaAcuPoint: MetaHuman-Generated Synthetic Forearm Data [Data set]. In MetaAcuPoint: MetaHuman-Generated Synthetic Data for Hand Acupoint Localization (v1.0). Zenodo. https://doi.org/10.5281/zenodo.17713204Dataset ContentsThe dataset consists of the following components:1. original_RGBA folder containing the 900 full-resolution RGB images generated under controlled lighting and consistent top-down imaging geometry.These images serve as the baseline dataset for training or benchmarking high-precision localization models.2. resized_RGBA folder containing the 900 resized versions of the original images. Resizing was performed to match input constraints of deep learning frameworks or to standardize spatial dimensions for downstream analysis. All resized images maintain the original aspect ratio and image quality suitable for annotation.3. annotation_Original_RGB.csvA comma-separated annotation file providing pixel-level labels for keypoints or region-of-interest coordinates corresponding to each image in the original_RGB folder.Each row contains:Image filenameX, Y coordinate values for the designated keypoints4. annotation_resized_RGB.jsonA JSON-formatted annotation file corresponding to the resized_RGB images. It follows a COCO-style schema, containing:Image metadata (filename, height, width)Keypoint annotationsSegmentation or bounding box fieldsCategory definitionsThis format is directly compatible with popular computer vision libraries such as MMPose and COCO API.5. avatar_description.xlsxAn Excel file containing descriptive information about the metahuman avatars included in the dataset. This includes demographic information, avatar attributes, and complementary metadata.Dataset Folder Structure (Unzip dataset_MetaAcuPoint.zip)dataset_MetaAcuPoint/│├── original_RGB/│   ├── img_0001.png│   ├── img_0002.png│   └── ...│├── resized_RGB/│   ├── img_0001.png│   ├── img_0002.png│   └── ...│├── annotation_Original_RGB.csv├── annotation_resized_RGB.json├── avatar_description.xlsx└── README.mdImage Naming ConventionEach image in the dataset follows a structured naming format: [AvatarID]arm[HandSide][FrameNumber].png | [XXX]arm[Y][Z].pngNote:AvatarID (XXX)A three-digit identifier ranging from 101 to 130Example: 101, 115, 130HandSide (X)Indicates which hand is shown    1 → Right hand    2 → Left handFrameNumber (Z)Frame index from 1 to 15, corresponding to the sequence extracted from the MetaHuman animation.