Dataset Overview
Definition：Dermatology Multimodal Datasetis provided byChangsha Langhui Information Technology Co., Ltd.Quality dermal data integrated. Covers clinical images, skin-dermal photographs and pathology 10+ diagnostic reports, creating image-clinical-pathological data links at three levels, and providing core training language for skin-disease AI screening and diagnosis.
| Dataset Name | Dermatology Multimodal Dataset |
| Total Data Volume | A lot of high-quality clinical images. |
| Data Type | Clinical Image + Structured Diagnostic Text |
| Data Source | Compliance |
| License Type | Authorization Complete authorization documents are attached |
| Core Features | 10+ pathological work check-ups, age of patients, skin damage, preliminary clinical diagnosis |
| Covered Disease Categories | Wet, silver crumbs, acne, measles, suspected rash, white flu, fat-spill, etc. |
Core Data Fields
Each piece of data contains complete structured fields that cover everything from raw data to specialized diagnostics.
| Field Category | Field Count | Typical Fields |
|---|---|---|
| ClinicalImage | 6 | Clinical image of oscillation/non-oscillation, magnification multiplier, image resolution, collection equipment model |
| Clinical photos | 4 | Macro photographs of skin lesions, lesion site annotation, shooting distance and lighting conditions |
| Pathology Report | 10 | full pathology diagnosis (10+), tissue type, celerity, leaching depth (Breslow), filamental division rate |
| Patient Information | 5 | Age, gender, lesion site, disease duration, preliminary clinical diagnosis (all de-identified) |
| Tab label | 3 | Benign-Malignant Classification, ABCDE Feature Scoring, 7-Point Checklist Scoring |
Data Sample Showcase
The following is a de-identified structured JSON sample illustrating the typical data format.
{
"case_id": "DERM-2023-0004521",
"meta": {"age": 58,"gender": "male","lesion_site": "sample value","duration": "6sample"},
"dermoscopy": {"type": "sample value","features": ["sample","sample","sample"]},
"pathology": {"diagnosis": "Unstable angina","breslow_depth": "2.4mm","ulceration": true,"mitotic_rate": "4/mm²","tnm_stage": "T3aN0M0"},
"clinical_dx": "sample value","final_label": "sample value"
}AI Application Scenarios
Skin screening AI
Training on clinical imagery-based CNN/Transformer models to achieve automatic classification and benign identification of common skin diseases, applicable to large-scale skin cancer screening.
Early Detection of Malignant Melanoma
The 10+ pathological coverage data are used to train early detection models for melanoma, combining ABCCE characteristics and seven-point checklists to improve the accuracy of early diagnosis.
Quality assessment of clinical images
Multi-light, multi-equipment-type clinical images are used to train the automatic image quality assessment model to assist remote dermal consultations.
Remote dermatology AI
A wealth of clinical images and clinical data was used to train remote dermatology AI systems to address the shortage of dermal doctors at the grass-roots level.
Pathology-Image Correlation Analysis
Pathology 10+ and clinical image-related data are used to train pathological prediction models, providing pre-invasive pathological type prognosis prior to the operation.
Escular AI teaching
A wealth of cases and data is used in the AIS teaching system of dermatology, which assists inpatients and medical students in identifying typical and unusual skin disorders.
Frequently asked questions
Domain expert network
Covering six major disciplines: mathematics, physics, chemistry, biology, earth sciences, and computer science
Academic credential review, professional competency assessment, data annotation training, and continuous quality monitoring
AI pre-annotation + expert review, 300% efficiency improvement, 99.5% accuracy
Need for customised medical data programs?
Changsha Langhui Information Technology Co., Ltd. The DataAssetsAPI platform is dedicated to providing high-quality, compliant medical data assets for AI companies and scientific research institutions. Cooperate with 50 + hospitals, including Fudan 100 type hospitals.
Data Sample Preview · Customized Data Solutions · Professional Technical Support