Document Type

Article

Source of Publication

IEEE Access

Publication Date

1-1-2021

Abstract

The biggest fear when deploying machine learning models to the real world is their ability to handle the new data. This problem is significant especially in medicine, where models trained on rich high-quality data extracted from large hospitals do not scale to small regional hospitals. One of the clinical challenges addressed in this work is magnetic resonance image generalization for improved visualization and diagnosis of hip abnormalities such as femoroacetabular impingement and dysplasia. Domain Generalization (DG) is a field in machine learning that tries to solve the model’s dependency on the training data by leveraging many related but different data sources. We present a new method for DG that is both efficient and fast, unlike the most current state of art methods, which add a substantial computational burden making it hard to fine-tune. Our model trains an autoencoder setting on top of the classifier, but the encoder is trained on the adversarial reconstruction loss forcing it to forget style information while extracting features useful for classification. Our approach aims to force the encoder to generate domain-invariant representations that are still category informative by pushing it in both directions. Our method has proven universal and was validated on four different benchmarks for domain generalization, outperforming state of the art on RMNIST, VLCS and IXMAS with a 0.70% increase in accuracy and providing comparable results on PACS with a 0.02% difference. Our method was also evaluated for unsupervised domain adaptation and has shown to be quite an effective method against over-fitting.

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Disciplines

Computer Sciences

Keywords

Computational modeling, Computer Vision, Data models, Deep Learning, Deep learning, Domain Adaptation, Domain Generalization, Feature extraction, Image reconstruction, Training, Transfer Learning, Transfer learning

Scopus ID

85103204777

Indexed in Scopus

yes

Open Access

yes

Open Access Type

Gold: This publication is openly available in an open access journal/series

Share

COinS