Skin Lesion Recognition with Class-Hierarchy Regularized Hyperbolic Embeddings

• Published in: Medical Image Computing and Computer Assisted Intervention - MICCAI 2022 Vol. 13433; pp. 594 - 603
• Main Authors: Yu, Zhen, Nguyen, Toan, Gal, Yaniv, Ju, Lie, Chandra, Shekhar S., Zhang, Lei, Bonnington, Paul, Mar, Victoria, Wang, Zhiyong, Ge, Zongyuan
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer 2022; Springer Nature Switzerland
• Series: Lecture Notes in Computer Science
• Subjects: Class hierarchy; Deep learning; Hyperbolic geometry; Skin lesion recognition
• ISBN: 3031164369; 9783031164361
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-031-16437-8_57

In practice, many medical datasets have an underlying taxonomy defined over the disease label space. However, existing classification algorithms for medical diagnoses often assume semantically independent labels. In this study, we aim to leverage class hierarchy with deep learning algorithms for more accurate and reliable skin lesion recognition. We propose a hyperbolic network to jointly learn image embeddings and class prototypes. The hyperbola provably provides a space for modeling hierarchical relations better than Euclidean geometry. Meanwhile, we restrict the distribution of hyperbolic prototypes with a distance matrix which is encoded from the class hierarchy. Accordingly, the learned prototypes preserve the semantic class relations in the embedding space and we can predict label of an image by assigning its feature to the nearest hyperbolic class prototype. We use an in-house skin lesion dataset which consists of ∼\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sim $$\end{document}230k dermoscopic images on 65 skin diseases to verify our method. Extensive experiments provide evidence that our model can achieve higher accuracy with less severe classification errors compared to that of models without considering class relations.