Skip to content

Latest commit

 

History

History
141 lines (104 loc) · 8.27 KB

PALM.md

File metadata and controls

141 lines (104 loc) · 8.27 KB

PALM: Open Fundus Photograph Dataset with Pathologic Myopia Recognition and Anatomical Structure Annotation

Dataset Information

Pathological myopia (PM) is a common blinding retinal degenerative condition among individuals with high myopia. Early screening for this condition can mitigate the damage caused by related fundus pathologies, thereby preventing vision loss. Automated diagnostic tools based on artificial intelligence can assist clinicians in identifying disease indicators or screening large populations using color fundus photographs as input, benefiting this process.

This paper provides insights into PALM, an open fundus imaging dataset designed for pathological myopia identification and anatomical structure annotation. Our database consists of 1,200 images, each labeled with pathological myopia categories, along with manual annotations for the optic disc, macular location, and lesions such as patchy atrophy (including peripapillary atrophy) and retinal detachment. Furthermore, this paper details the annotation process used to construct the database, the quality and characteristics of the samples, and additional usage instructions.

Figure 1. Examples of common retinal lesions in pathological myopia (PM) cases include: (a)Peripapillary atrophy, occurring near the optic disc; (b)Reticular retina, where prominent large choroidal vessels can be observed in the posterior pole of the fundus; (c)Macular hemorrhage, primarily along the fissure itself and at or near the macular center; (d)Retinal atrophy, caused by the migration of degenerated retinal pigment epithelial cells to the inner retinal layers, leading to lesions in and around the affected areas.

Dataset Meta Information

Dimensions Modality Task Type Anatomical Area Number of Categories Data Volume File Format
2D Fundus Photography Classification, Segmentation Retina 5 or 2 1200 JPG

Resolution Details

Dataset Statistics size
min (1444, 1444)
median (2124, 2056)
max (2124, 2056)

Label Information Statistics

Set Num. PM/Non-PM With/Without OD With/Without Fovea With/Without Detachment With/Without Atrophy Photo Centering (OD/Fovea/Midpoint of OD and Fovea) Device (Zeiss/Canon)
Training 400 213/187 381/19 397/3 12/388 311/89 42/258/100 350/50
Validation 400 211/189 379/21 397/3 6/394 271/129 43/258/99 344/56
Testing 400 213/187 384/16 398/2 6/394 288/112 38/284/78 353/47
Total 1200 637/563 1144/56 1192/8 24/1176 870/330 123/800/277 1047/153

Explanation of Columns:

  • Set: Dataset split into training, validation, and testing subsets.
  • Num.: Number of images in each set.
  • PM/Non-PM: Images labeled as Pathological Myopia (PM) or Non-Pathological Myopia (Non-PM).
  • With/Without OD: Images with or without Optic Disc (OD).
  • With/Without Fovea: Images with or without the presence of the fovea.
  • With/Without Detachment: Images with or without retinal detachment.
  • With/Without Atrophy: Images with or without retinal atrophy.
  • Photo Centering: Centering of the photo based on OD, Fovea, or the midpoint between OD and Fovea.
  • Device: Equipment used to capture images (Zeiss or Canon).

Visualization

Figure 1. Examples of common retinal lesions in cases of pathological myopia (PM) include: (a) Peripapillary atrophy; (b) Reticular retina; (c) Macular hemorrhage; (d) Retinal atrophy; (e) Retinal detachment; (f) Vitreous opacity. All images are sourced from the PALM training samples.

Figure 2. Examples of annotation interfaces used by experts for (a) non-pathological myopia samples and (b) pathological myopia samples. (a1) and (b1): Original input images; (a2) and (b2): Manual annotations.

File Structure

PALM
│
├── Training
│   ├── hippocampus_001.nii.gz
│   │   └── Images
│   │       ├── ...
│   │   └── Disc Masks
│   │       ├── ...
│   │   └── Lesion Masks
│   │       ├── Atrophy
│   │           ├── ...
│   │       ├── Detachment
│   │           ├── ...
│   │
│   └── Classification Labels
│   └── Fovea Localization
│   └── Supplementary Information
│
├── Validation
│   ├── ...
│   └── ...
├── Testing
│   ├── ...
│   └── ...

Authors and Institutions

Huihui Fang (South China University of Technology, Guangzhou, China; Pazhou Lab., Guangzhou, China)

Fei Li (State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China)

Junde Wu (National University of Singapore, Singapore, Singapore)

Huazhu Fu (Institute of High Performance Computing, Agency for Science, Technology and Research, Singapore, Singapore)

Xu Sun (Pazhou Lab., Guangzhou, China)

José Ignacio Orlando (Yatiris Group, PLADEMA Institute, CONICET, UNICEN, Tandil, Argentina)

Hrvoje Bogunović (Christian Doppler Lab for Artificial Intelligence in Retina, Department of Ophthalmology and Optometry, Medical University of Vienna, Vienna, Austria)

Xiulan Zhang (State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China)

Yanwu Xu (South China University of Technology, Guangzhou, China; Pazhou Lab., Guangzhou, China)

Source Information

Official Website: https://www.nature.com/articles/s41597-024-02911-2?_gl=1*16u2cf9*_up*MQ..&gclid=CjwKCAjw9eO3BhBNEiwAoc0-jXF2c7y1hinoDc71AuwyTwBqCuBlioxEwhfIUDUuFRPDtXrGn6S7HhoC4IMQAvD_BwE

Download Link: https://palm.grand-challenge.org/SemifinalLeaderboard/

Article Address: https://www.nature.com/articles/s41597-024-02911-2?_gl=1*16u2cf9*_up*MQ..&gclid=CjwKCAjw9eO3BhBNEiwAoc0-jXF2c7y1hinoDc71AuwyTwBqCuBlioxEwhfIUDUuFRPDtXrGn6S7HhoC4IMQAvD_BwE

Publication Date: 2024-01

Citation

@article{fang2024open,
  title={Open fundus photograph dataset with pathologic myopia recognition and anatomical structure annotation},
  author={Fang, Huihui and Li, Fei and Wu, Junde and Fu, Huazhu and Sun, Xu and Orlando, Jos{\'e} Ignacio and Bogunovi{\'c}, Hrvoje and Zhang, Xiulan and Xu, Yanwu},
  journal={Scientific Data},
  volume={11},
  number={1},
  pages={99},
  year={2024},
  publisher={Nature Publishing Group UK London}
}

Original introduction article is here.