[Paper Review] An Image Dataset of Common Skin Diseases of Bangladesh and Benchmarking Performance with Machine Learning Models

Md. Sazzad Hossain, Saiful Islam+4 more|arXiv (Cornell University)|Mar 26, 2026

Cutaneous Melanoma Detection and Management0 citations

TL;DR

The paper presents a publicly available image dataset of five common Bangladeshi skin diseases and benchmarks multiple ML/DL models on it. It also discusses dataset collection and regional relevance for global dermatology applications.

ABSTRACT

Skin diseases are a major public health concern worldwide, and their detection is often challenging without access to dermatological expertise. In countries like Bangladesh, which is highly populated, the number of qualified skin specialists and diagnostic instruments is insufficient to meet the demand. Due to the lack of proper detection and treatment of skin diseases, that may lead to severe health consequences including death. Common properties of skin diseases are, changing the color, texture, and pattern of skin and in this era of artificial intelligence and machine learning, we are able to detect skin diseases by using image processing and computer vision techniques. In response to this challenge, we develop a publicly available dataset focused on common skin disease detection using machine learning techniques. We focus on five prevalent skin diseases in Bangladesh: Contact Dermatitis, Vitiligo, Eczema, Scabies, and Tinea Ringworm. The dataset consists of 1612 images (of which, 250 are distinct while others are augmented), collected directly from patients at the outpatient department of Faridpur Medical College, Faridpur, Bangladesh. The data comprises of 302, 381, 301, 316, and 312 images of Dermatitis, Eczema, Scabies, Tinea Ringworm, and Vitiligo, respectively. Although the data are collected regionally, the selected diseases are common across many countries especially in South Asia, making the dataset potentially valuable for global applications in machine learning-based dermatology. We also apply several machine learning and deep learning models on the dataset and report classification performance. We expect that this research would garner attention from machine learning and deep learning researchers and practitioners working in the field of automated disease diagnosis.

Motivation & Objective

• Motivate automated detection of skin diseases in Bangladesh where dermatological expertise is limited.
• Create a publicly available image dataset focusing on prevalent skin diseases.
• Benchmark a range of machine learning and deep learning models on the dataset to establish baseline performance.
• Highlight the potential global relevance of the dataset for ML-based dermatology research.

Proposed method

• Assemble a regional skin disease image dataset from outpatient patients at a Bangladeshi medical college hospital.
• Categorize images into five diseases: Contact Dermatitis, Vitiligo, Eczema, Scabies, and Tinea Ringworm.
• Provide a dataset composition including augmented versus distinct images (1612 total, with 250 distinct).
• Apply several machine learning and deep learning models to perform disease classification and report performance.
• Discuss the potential applicability of the dataset for ML-based dermatology research in low-resource settings.

Experimental results

Research questions

• RQ1 Can a publicly available image dataset of common Bangladeshi skin diseases support ML-based detection?
• RQ2 How do various ML/DL models perform on this dataset across five disease classes?
• RQ3 What is the impact of data augmentation on model performance for skin disease classification?
• RQ4 Is the dataset potentially valuable for global dermatology applications beyond the regional context?

Key findings

• Dataset contains 1612 images across five diseases: Dermatitis (302), Eczema (381), Scabies (301), Tinea Ringworm (316), Vitiligo (312).
• 250 images are distinct; remaining are augmented to expand the dataset.
• A range of machine learning and deep learning models were benchmarked on the dataset, with reported classification performance (specific metrics not provided in the abstract).
• The authors suggest the dataset’s regional collection may still be valuable for global applications in ML-based dermatology.