Bias Assessment and Data Drift Detection in Medical Image Analysis: A Survey

September 2024 in “ arXiv (Cornell University) ”

Andrea Prenner, Bernhard Kainz, Kainz, Bernhard

Studysummary This study reviews methods to ensure the reliability of machine learning models in medical imaging, focusing on bias detection, data drift assessment, and accuracy estimation without ground truth labels to enhance integration into clinical settings. Our plain-language summary of this paper — not a Tressless recommendation.

This survey reviews methods to ensure the reliability of machine learning models in medical imaging analysis, focusing on bias assessment and data drift detection. It categorizes techniques for evaluating models' inner workings, particularly in disease classification, and discusses methods for estimating classifier accuracy without ground truth labels. The goal is to enhance the trustworthiness and integration of these models into clinical settings by maintaining consistent prediction performance over time.

View this study on arxiv.org →

Discuss this study in the Community →

Similar Research

5 / 1000+ results

research Skin image analysis for detection and quantitative assessment of dermatitis, vitiligo and alopecia areata lesions: a systematic literature review

5 citations , January 2025 in “BMC Medical Informatics and Decision Making”

This review examines the use of computer vision techniques, specifically deep learning architectures and image processing algorithms, for detecting and assessing skin conditions like vitiligo and dermatitis, and highlights the need for disease-specific datasets to improve automated diagnostic tools in dermatology.

research Results in a ‘Flash’: Recognizing Misleading Clinical Photography

May 2026 in “Indian Dermatology Online Journal”

This study highlights the prevalence of photographic bias in before-and-after images in dermatology, emphasizing how factors like lighting and camera angles can mislead patients and influence expectations, often leading to dissatisfaction and unrealistic aesthetic standards.

research A Comprehensive Survey on Vitiligo Detection Using Deep Learning

January 2026 in “ITM Web of Conferences”

This review examines the current state of automated vitiligo detection systems, noting a lack of large, diverse datasets and consistent imaging conditions, while comparing traditional and modern machine learning approaches to improve reliability and applicability.

research Closing the AI generalization gap by adjusting for dermatology condition distribution differences across clinical settings

February 2024 in “arXiv (Cornell University)”

In this study, researchers found that differences in skin condition distribution are the main source of errors when AI algorithms classify dermatological conditions from new, previously unseen sources, and proposed steps to improve their generalizability based on available information.

research Big Self-Supervised Models Advance Medical Image Classification

January 2021 in “arXiv (Cornell University)”

This study found that self-supervised pretraining significantly improves accuracy in medical image classifiers for dermatology and chest X-ray tasks, outperforming supervised baselines and showing robustness to distribution shifts with limited labeled data.