Enhanced Dermatology Image Quality Assessment via Cross-Domain Training

2025-06-19

AI Summary

Poor image quality severely undermines diagnostic reliability in teledermatology, while existing dermatological image quality assessment (IQA) methods suffer from limited training data, narrow domain coverage, and failure to leverage advances in general-purpose IQA. To address these limitations, we propose a cross-domain joint training framework. We first introduce Legit.Health-DIQA-Artificial—a novel, dermatology-specific synthetic IQA dataset—and jointly train on it alongside large-scale general IQA benchmarks (e.g., LIVE, TID2013). Our approach integrates human subjective quality annotations with deep transfer learning. By bridging domain gaps and enriching supervision signals, the method overcomes the small-sample bottleneck, substantially improving model generalizability and robustness. Extensive evaluation across multiple domains demonstrates state-of-the-art accuracy in dermatological IQA. This work establishes a reliable, scalable foundation for image quality screening in remote dermatological diagnosis and triage.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Knowledge Representation and Reasoning: Qualitative Reasoning
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Web data quality in the era of algorithmically-generated content
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Teledermatology has become a widely accepted communication method in daily clinical practice, enabling remote care while showing strong agreement with in-person visits. Poor image quality remains an unsolved problem in teledermatology and is a major concern to practitioners, as bad-quality images reduce the usefulness of the remote consultation process. However, research on Image Quality Assessment (IQA) in dermatology is sparse, and does not leverage the latest advances in non-dermatology IQA, such as using larger image databases with ratings from large groups of human observers. In this work, we propose cross-domain training of IQA models, combining dermatology and non-dermatology IQA datasets. For this purpose, we created a novel dermatology IQA database, Legit.Health-DIQA-Artificial, using dermatology images from several sources and having them annotated by a group of human observers. We demonstrate that cross-domain training yields optimal performance across domains and overcomes one of the biggest limitations in dermatology IQA, which is the small scale of data, and leads to models trained on a larger pool of image distortions, resulting in a better management of image quality in the teledermatology process.

Problem

• Improving dermatology image quality assessment via cross-domain training
• Addressing limited data in dermatology IQA with combined datasets
• Enhancing teledermatology by managing diverse image distortions effectively

Innovation

• Cross-domain training for dermatology IQA
• Combining dermatology and non-dermatology datasets
• Leveraging human-annotated image distortions