Datasets:

MaybeRichard
/

zju-eye-pretrain

cohort stringclasses 1 value	study_id stringlengths 16 16	patient_hash stringlengths 16 16	visit_date stringdate 2018-08-04 00:00:00 2025-12-15 00:00:00	eye stringclasses 2 values	device_vendor stringclasses 1 value	device_model stringclasses 1 value	device_serial_hash stringclasses 8 values	device_software_version stringclasses 5 values	hospital_domain stringclasses 1 value	ethnicity stringclasses 1 value	image_quality_score float64 58.2 94.7 ⌀	image_quality_band stringclasses 5 values	diagnosis_group listlengths 0 0	lesion_tags listlengths 0 0	lesion_location listlengths 0 0	layer_involvement listlengths 0 0	severity stringclasses 1 value	diagnosis_source stringclasses 1 value	label_confidence null	schema_version stringclasses 1 value	image_id stringlengths 48 57	file_path stringlengths 32 41	file_format stringclasses 2 values	modality stringclasses 3 values	anatomy stringclasses 3 values	device_technology stringclasses 3 values	scan_protocol stringclasses 4 values	scan_x_mm float64 6 12 ⌀	bscan_index float64 0 319 ⌀	image_height_px int64 480 1.93k	image_width_px int64 320 2.58k	axial_resolution_um float64 2.55 2.66 ⌀	has_segmentation bool 2 classes	n_layers_visible int64 0 10	fovea_x_norm float64 0.3 0.7 ⌀	crt_um float64 82.4 911 ⌀	choroid_thickness_um float64 28.1 733 ⌀	oct_footprint_bbox_fundus listlengths 4 4 ⌀	oct_footprint_bbox_slo listlengths 4 4 ⌀	is_valid bool 1 class	image unknown
shanghai_drioct_triton	0008410619c9f886	0008410619c9f886	2021-02-03	OS	topcon	dri_oct_triton	970026	10.1.6.49687	shanghai_zhongshan_v1	Asian	82.095947	good	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_0008410619c9f886_fundus_color	00/0008410619c9f886/fundus_color.jpg	jpg	fundus_color	macula	fundus_camera	single_shot	null	null	1,934	2,576	null	false	0	null	null	null	[ 860, 522, 1748, 1408 ]	null	true	"/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAIBAQEBAQIBAQECAgICAgQDAgICAgUEBAMEBgUGBgYFBgYGBwkIBgcJBwYGCAsICQo(...TRUNCATED)
shanghai_drioct_triton	0008410619c9f886	0008410619c9f886	2021-02-03	OS	topcon	dri_oct_triton	970026	10.1.6.49687	shanghai_zhongshan_v1	Asian	82.095947	good	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_0008410619c9f886_slo_gray	00/0008410619c9f886/slo_gray.png	png	slo_gray	macula	slo	single_shot	null	null	480	640	null	false	0	null	null	null	null	[ 213, 130, 431, 350 ]	true	"iVBORw0KGgoAAAANSUhEUgAAAoAAAAHgCAAAAAAQuoM4AAEAAElEQVR42sT9XXob2bIsCJqZrwAl5e/epwZUg+u3nlmPpu6tW31(...TRUNCATED)
shanghai_drioct_triton	0008410619c9f886	0008410619c9f886	2021-02-03	OS	topcon	dri_oct_triton	970026	10.1.6.49687	shanghai_zhongshan_v1	Asian	78.058159	moderate	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_0008410619c9f886_subscan_bscan_000	00/0008410619c9f886/subscan_bscan/000.png	png	oct_bscan	optic_disc	ss_oct	subscan_line	12	0	992	1,024	2.621349	true	10	null	null	null	[ 404, 976, 2188, 958 ]	[ 103, 240, 541, 240 ]	true	"iVBORw0KGgoAAAANSUhEUgAABAAAAAPgCAAAAACJlVFTAAEAAElEQVR42oz92ZIkSbIsBjKLqkdkdTcAovmDoQt86PzwEOYsXRn(...TRUNCATED)
shanghai_drioct_triton	0008410619c9f886	0008410619c9f886	2021-02-03	OS	topcon	dri_oct_triton	970026	10.1.6.49687	shanghai_zhongshan_v1	Asian	78.058159	moderate	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_0008410619c9f886_subscan_slo	00/0008410619c9f886/subscan_slo.png	png	slo_gray	optic_disc	slo	subscan_single_shot	null	null	480	640	null	false	0	null	null	null	null	[ 103, 240, 541, 240 ]	true	"iVBORw0KGgoAAAANSUhEUgAAAoAAAAHgCAAAAAAQuoM4AAEAAElEQVR42sz96YIjSdIkiImImiPyqsyqnn1j/uarklzODr+phJs(...TRUNCATED)
shanghai_drioct_triton	00035a34a0979571	00035a34a0979571	2018-08-04	OS	topcon	dri_oct_triton	980136	10.1.3.44760	shanghai_zhongshan_v1	Asian	85.328484	good	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_00035a34a0979571_bscan_radial_000	00/00035a34a0979571/bscan_radial/000.png	png	oct_bscan	macula	ss_oct	radial_12	6	0	992	1,024	2.609594	true	10	0.516	216.6	224.4	null	null	true	"iVBORw0KGgoAAAANSUhEUgAABAAAAAPgCAAAAACJlVFTAAEAAElEQVR42lz927pl2XGkibn5WHHITCTAQ7GOTYAl3UuXev97vYF(...TRUNCATED)
shanghai_drioct_triton	00035a34a0979571	00035a34a0979571	2018-08-04	OS	topcon	dri_oct_triton	980136	10.1.3.44760	shanghai_zhongshan_v1	Asian	85.328484	good	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_00035a34a0979571_bscan_radial_001	00/00035a34a0979571/bscan_radial/001.png	png	oct_bscan	macula	ss_oct	radial_12	6	1	992	1,024	2.609594	true	10	0.501	221.8	227	null	null	true	"iVBORw0KGgoAAAANSUhEUgAABAAAAAPgCAAAAACJlVFTAAEAAElEQVR42mT927ZmSZKrhxnMV2TxcKGha7KrW8+h938iDYldlTE(...TRUNCATED)
shanghai_drioct_triton	00035a34a0979571	00035a34a0979571	2018-08-04	OS	topcon	dri_oct_triton	980136	10.1.3.44760	shanghai_zhongshan_v1	Asian	85.328484	good	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_00035a34a0979571_bscan_radial_002	00/00035a34a0979571/bscan_radial/002.png	png	oct_bscan	macula	ss_oct	radial_12	6	2	992	1,024	2.609594	true	10	0.497	221.8	221.8	null	null	true	"iVBORw0KGgoAAAANSUhEUgAABAAAAAPgCAAAAACJlVFTAAEAAElEQVR42my9686sWXKcl5Grvt3dQ0q0IMOGIU7PiPZd+P5vwr9(...TRUNCATED)
shanghai_drioct_triton	00035a34a0979571	00035a34a0979571	2018-08-04	OS	topcon	dri_oct_triton	980136	10.1.3.44760	shanghai_zhongshan_v1	Asian	85.328484	good	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_00035a34a0979571_bscan_radial_003	00/00035a34a0979571/bscan_radial/003.png	png	oct_bscan	macula	ss_oct	radial_12	6	3	992	1,024	2.609594	true	10	0.488	227	214	null	null	true	"iVBORw0KGgoAAAANSUhEUgAABAAAAAPgCAAAAACJlVFTAAEAAElEQVR42lz93bJlW3KciblHzJ2nUABIqlvWLZICiuw205XeQO/(...TRUNCATED)
shanghai_drioct_triton	00035a34a0979571	00035a34a0979571	2018-08-04	OS	topcon	dri_oct_triton	980136	10.1.3.44760	shanghai_zhongshan_v1	Asian	85.328484	good	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_00035a34a0979571_bscan_radial_004	00/00035a34a0979571/bscan_radial/004.png	png	oct_bscan	macula	ss_oct	radial_12	6	4	992	1,024	2.609594	true	10	0.495	216.6	216.6	null	null	true	"iVBORw0KGgoAAAANSUhEUgAABAAAAAPgCAAAAACJlVFTAAEAAElEQVR42lT9W7NtWZKch7lHzJ2X6guuBCgRXd0kzfQf9KD//yy(...TRUNCATED)
shanghai_drioct_triton	00035a34a0979571	00035a34a0979571	2018-08-04	OS	topcon	dri_oct_triton	980136	10.1.3.44760	shanghai_zhongshan_v1	Asian	85.328484	good	[]	[]	[]	[]	unknown	none	null	v1	shanghai_drioct_triton_00035a34a0979571_bscan_radial_005	00/00035a34a0979571/bscan_radial/005.png	png	oct_bscan	macula	ss_oct	radial_12	6	5	992	1,024	2.609594	true	10	0.453	206.2	224.4	null	null	true	"iVBORw0KGgoAAAANSUhEUgAABAAAAAPgCAAAAACJlVFTAAEAAElEQVR42lz927Ks2XKcibnHyLlqb5wIQLRWGwkQbFJ6CJnp/R9(...TRUNCATED)

End of preview. Expand in Data Studio

ZJU Eye-Pretrain Dataset

Unified multi-source ophthalmological imaging dataset for foundation model pretraining and downstream tasks.

1.1M images spanning 26 cohorts with a strict 41-column unified manifest schema.

Composition

Source	Images	Modalities	Cohorts
Private Shanghai DRI OCT Triton (SS-OCT)	419,042	oct_bscan + fundus_color + slo_gray	1
Public Fundus	198,629	fundus_color (+ GAMMA OCT)	6
Public OCT	488,705	oct_bscan	19
Total	1,106,376		26

See DATASET_OVERVIEW.md for full details per cohort (devices, regions, masks, demographics).

Quick Start

from datasets import load_dataset

# Load everything (1.1M images)
ds = load_dataset("mayberichard/zju-eye-pretrain", streaming=True)

# Load one batch
ds = load_dataset("mayberichard/zju-eye-pretrain", "public_oct")

# Load one cohort
ds = load_dataset("mayberichard/zju-eye-pretrain", "kermany")

# Each row:
#   image: PIL.Image
#   {layer/lesion/vessel/disc_cup}_mask: PIL.Image or None
#   image_id, study_id, patient_hash, modality, anatomy, severity, ...

Schema (41-column manifest, identical across all batches)

cohort, study_id, patient_hash, visit_date, eye,
device_vendor, device_model, device_serial_hash, device_software_version,
hospital_domain, ethnicity,
image_quality_score, image_quality_band,
diagnosis_group, lesion_tags, lesion_location, layer_involvement, severity,
diagnosis_source, label_confidence, schema_version,
image_id, file_path, file_format,
modality, anatomy, device_technology, scan_protocol,
scan_x_mm, bscan_index,
image_height_px, image_width_px, axial_resolution_um,
has_segmentation, n_layers_visible,
fovea_x_norm, crt_um, choroid_thickness_um,
oct_footprint_bbox_fundus, oct_footprint_bbox_slo,
is_valid

Plus per-image image bytes and per-cohort mask columns.

Captions

Each image has 5 captions (4 L1 variants + 1 L3 derived). Total 5.5M captions in captions/.

from datasets import load_dataset
caps = load_dataset("mayberichard/zju-eye-pretrain", "captions_oct")
# join on image_id with the images config

Licensing

This dataset aggregates multiple sources with mixed licenses. See LICENSE for per-cohort license terms. Users are responsible for compliance with the original license of each cohort.

Private Shanghai Topcon data is included for research convenience. Commercial use is prohibited.

Citation

If you use this dataset, please cite the original source for each cohort used (see DATASET_OVERVIEW.md).

Versioning & Updates

This dataset supports incremental updates. New cohorts can be added without touching existing data via additional shards in data/<batch>/. Schema migrations preserve old *_v1.parquet alongside new versions.

Downloads last month: -