Presentation
15 February 2021 Overcoming the "catastrophic forgetting" effect in transfer learning to achieve vendor independent performance for the COVID-19 pneumonia classification task using chest x-ray radiographs
Ran Zhang, Guang-Hong Chen
Author Affiliations +
Abstract
In this work, we studied the catastrophic forgetting effect in transfer learning for the development of a deep learning model to perform COVID-19 pneumonia classification using chest x-ray radiographs. We demonstrated vendor differences using a large multi-vendor chest x-ray dataset for COVID-19 classification and investigated its impact on the generalizability of the model. We showed that the model trained from one particular vendor may not generalize well to other vendors. While both vendor-based fine-tuning and joint training can improve the performance on the new vendor test set, the first approach may cause performance degradation on the previous vendor test set.
Conference Presentation
© (2021) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Ran Zhang and Guang-Hong Chen "Overcoming the "catastrophic forgetting" effect in transfer learning to achieve vendor independent performance for the COVID-19 pneumonia classification task using chest x-ray radiographs", Proc. SPIE 11597, Medical Imaging 2021: Computer-Aided Diagnosis, 115971R (15 February 2021); https://doi.org/10.1117/12.2581088
Advertisement
Advertisement
KEYWORDS
Data modeling

Chest imaging

Radiography

Performance modeling

Imaging systems

Image processing

Neural networks

Back to Top