Dermatological Conditions Image Dataset for Skin Lesion Classification
Description
This dataset contains labeled images of various dermatological conditions intended for research and development of computer vision models for skin disease classification. The dataset is organized into folders, each representing a distinct skin condition. Images were classified into different folders based on visible clinical features associated with each disease. Folder Structure / Classes: The dataset includes the following classes: Acne – Images showing comedones, papules, pustules, and nodules characteristic of acne vulgaris. Hyperpigmentation – Images of skin areas exhibiting excess melanin production, including melasma and post-inflammatory hyperpigmentation. Nail Psoriasis – Images showing psoriatic nail changes such as pitting, onycholysis, and subungual hyperkeratosis. SJS-TEN (Stevens-Johnson Syndrome / Toxic Epidermal Necrolysis) – Clinical photographs depicting severe mucocutaneous reactions with epidermal detachment and erythematous lesions. Vitiligo – Images illustrating depigmented patches caused by loss of melanocytes. Data Format: Each folder contains JPEG/PNG image files. Images vary in resolution and lighting conditions. File naming conventions follow a consistent pattern for easy data handling. Intended Use: This dataset is suitable for: Image classification tasks in dermatology. Deep learning and machine learning model training. Computer-aided diagnostic system development. Research in medical image analysis and skin disease detection. Ethical Considerations: All images were anonymized to remove identifiable information. The dataset is shared strictly for academic and research purposes. Users must comply with ethical standards and data usage policies when using this dataset.
Files
Institutions
- University of North Texas
- Daffodil International University
- Jagannath University