An analysis of data variation and bias in image-based dermatological datasets for machine learning classification

TL;DR

This paper investigates the differences between dermoscopic and clinical skin image datasets, analyzing how data variation and bias affect machine learning classification performance in dermatology, and proposes methods to mitigate these effects.

Contribution

It provides an analysis of dataset distribution gaps and proposes strategies to combine divergent data sources to improve model accuracy in clinical dermatology applications.

Findings

Dataset variations significantly impact model performance.

Transfer learning can be affected by dataset bias and size.

Combining data sources reduces prediction errors.

Abstract

AI algorithms have become valuable in aiding professionals in healthcare. The increasing confidence obtained by these models is helpful in critical decision demands. In clinical dermatology, classification models can detect malignant lesions on patients' skin using only RGB images as input. However, most learning-based methods employ data acquired from dermoscopic datasets on training, which are large and validated by a gold standard. Clinical models aim to deal with classification on users' smartphone cameras that do not contain the corresponding resolution provided by dermoscopy. Also, clinical applications bring new challenges. It can contain captures from uncontrolled environments, skin tone variations, viewpoint changes, noises in data and labels, and unbalanced classes. A possible alternative would be to use transfer learning to deal with the clinical images. However, as the number of samples is low, it can cause degradations on the model's performance; the source distribution used in training differs from the test set. This work aims to evaluate the gap between dermoscopic and clinical samples and understand how the dataset variations impact training. It assesses the main differences between distributions that disturb the model's prediction. Finally, from experiments on different architectures, we argue how to combine the data from divergent distributions, decreasing the impact on the model's final accuracy.