<abstract xmlns="http://www.w3.org/1999/xhtml">

There has been an amplified focus on and benefit from the adoption of artificial intelligence (AI) in medical imaging applications. However, deep learning approaches involve training with massive amounts of annotated data in order to guarantee generalization and achieve high accuracies. Gathering and annotating large sets of training images require expertise which is both expensive and time-consuming, especially in the medical field. Furthermore, in health care systems where mistakes can have catastrophic consequences, there is a general mistrust in the black-box aspect of AI models. In this work, we focus on improving the performance of medical imaging applications when limited data is available while focusing on the interpretability aspect of the proposed AI model. This is achieved by employing a novel transfer learning framework, <italic>progressive transfer learning</italic>, an automated annotation technique and a correlation analysis experiment on the learned representations.
<italic>Progressive transfer learning</italic> helps jump-start the training of deep neural networks while improving the performance by gradually transferring knowledge from two source tasks into the target task. It is empirically tested on the wrist fracture detection application by first training a general radiology network <italic>RadiNet</italic> and using its weights to initialize <italic>RadiNetwrist</italic>, that is trained on wrist images to detect fractures. Experiments show that <italic>RadiNetwrist</italic> achieves an accuracy of 87% and an AUC ROC of 94% as opposed to 83% and 92% when it is pre-trained on the ImageNet dataset.
This improvement in performance is investigated within an <italic>explainable AI</italic> framework. More concretely, the learned deep representations of <italic>RadiNetwrist</italic> are compared to those learned by the baseline model by conducting a correlation analysis experiment. The results show that, when transfer learning is <italic>gradually</italic> applied, some features are learned earlier in the network. Moreover, the deep layers in the <italic>progressive transfer learning</italic> framework are shown to encode features that are not encountered when traditional transfer learning techniques are applied.
In addition to the empirical results, a clinical study is conducted and the performance of <italic>RadiNetwrist</italic> is compared to that of an expert radiologist. We found that <italic>RadiNetwrist</italic> exhibited similar performance to that of radiologists with more than 20 years of experience.
This motivates follow-up research to train on more data to feasibly surpass radiologists’ performance, and investigate the interpretability of AI models in the healthcare domain where the decision-making process needs to be credible and transparent.
</abstract>

There has been an amplified focus on and benefit from the adoption of artificial intelligence (AI) in medical imaging applications. However, deep learning approaches involve training with massive amounts of annotated data in order to guarantee generalization and achieve high accuracies. Gathering and annotating large sets of training images require expertise which is both expensive and time-consuming, especially in the medical field. Furthermore, in health care systems where mistakes can have catastrophic consequences, there is a general mistrust in the black-box aspect of AI models. In this work, we focus on improving the performance of medical imaging applications when limited data is available while focusing on the interpretability aspect of the proposed AI model. This is achieved by employing a novel transfer learning framework, progressive transfer learning, an automated annotation technique and a correlation analysis experiment on the learned representations.
Progressive transfer learning helps jump-start the training of deep neural networks while improving the performance by gradually transferring knowledge from two source tasks into the target task. It is empirically tested on the wrist fracture detection application by first training a general radiology network RadiNet and using its weights to initialize RadiNetwrist, that is trained on wrist images to detect fractures. Experiments show that RadiNetwrist achieves an accuracy of 87% and an AUC ROC of 94% as opposed to 83% and 92% when it is pre-trained on the ImageNet dataset.
This improvement in performance is investigated within an explainable AI framework. More concretely, the learned deep representations of RadiNetwrist are compared to those learned by the baseline model by conducting a correlation analysis experiment. The results show that, when transfer learning is gradually applied, some features are learned earlier in the network. Moreover, the deep layers in the progressive transfer learning framework are shown to encode features that are not encountered when traditional transfer learning techniques are applied.
In addition to the empirical results, a clinical study is conducted and the performance of RadiNetwrist is compared to that of an expert radiologist. We found that RadiNetwrist exhibited similar performance to that of radiologists with more than 20 years of experience.
This motivates follow-up research to train on more data to feasibly surpass radiologists’ performance, and investigate the interpretability of AI models in the healthcare domain where the decision-making process needs to be credible and transparent.

A Progressive and Cross-Domain Deep Transfer Learning Framework for Wrist Fracture Detection

Department of Electrical and Computer Engineering

Department of Health Professions, Medical Imaging Sciences, Faculty of Health Sciences

Department of Radiology, Faculty of Medicine

Journal of Artificial Intelligence and Soft Computing Research

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

There has been an amplified focus on and benefit from the adoption of artificial intelligence (AI) in medical imaging applications. However, deep learning...

A Progressive and Cross-Domain Deep Transfer Learning Framework for Wrist Fracture Detection

Published Online: Feb 23, 2022

Page range: 101 - 120

Received: May 24, 2021

Accepted: Aug 09, 2021

DOI: https://doi.org/10.2478/jaiscr-2022-0007

Keywords
deep learning, transfer learning, wrist fracture detection, medical informatics, progressive transfer learning

© 2022 Christophe Karam et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

A Progressive and Cross-Domain Deep Transfer Learning Framework for Wrist Fracture Detection

Published Online: Feb 23, 2022

Page range: 101 - 120

Received: May 24, 2021

Accepted: Aug 09, 2021

DOI: https://doi.org/10.2478/jaiscr-2022-0007

Keywordsdeep learning, transfer learning, wrist fracture detection, medical informatics, progressive transfer learning

© 2022 Christophe Karam et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Keywords
deep learning, transfer learning, wrist fracture detection, medical informatics, progressive transfer learning