A Deep Multi-task Learning Approach to Skin Lesion Classification
This is an incremental improvement for dermatological diagnosis by leveraging body site information to enhance skin lesion classification.
The paper tackled skin lesion classification by incorporating body location as additional context, using a deep multi-task learning framework to jointly optimize lesion and location classification, resulting in more robust performance compared to a single-task approach.
Skin lesion identification is a key step toward dermatological diagnosis. When describing a skin lesion, it is very important to note its body site distribution as many skin diseases commonly affect particular parts of the body. To exploit the correlation between skin lesions and their body site distributions, in this study, we investigate the possibility of improving skin lesion classification using the additional context information provided by body location. Specifically, we build a deep multi-task learning (MTL) framework to jointly optimize skin lesion classification and body location classification (the latter is used as an inductive bias). Our MTL framework uses the state-of-the-art ImageNet pretrained model with specialized loss functions for the two related tasks. Our experiments show that the proposed MTL based method performs more robustly than its standalone (single-task) counterpart.