Survey of Image Denoising Methods for Medical Image Classification
Abstract
Medical imaging devices, such as X-ray machines, inherently produce images that suffer from visual noise. Our objectives were to (i.) determine the effect of image denoising on a medical image classification task, and (ii.) determine if there exists a correlation between image denoising performance and medical image classification performance. We performed the medical image classification task on chest X-rays using the DenseNet-121 convolutional neural network (CNN) and used the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) metrics as the image denoising performance measures. We first found that different denoising methods can make a statistically significant difference in classification performance for select labels. We also found that denoising methods affect fine-tuned models more than randomly-initialized models and that fine-tuned models have significantly higher and more uniform performance than randomly-initialized models. Lastly, we found that there is no significant correlation between PSNR and SSIM values and classification performance for our task.
- Authors:
-
- University of Washington, Seattle
- ORNL
- Publication Date:
- Research Org.:
- Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
- Sponsoring Org.:
- USDOE
- OSTI Identifier:
- 1648905
- DOE Contract Number:
- AC05-00OR22725
- Resource Type:
- Conference
- Resource Relation:
- Conference: SPIE Medical Imaging 2020 - Houston, Texas, United States of America - 3/16/2020 4:00:00 AM-3/20/2020 4:00:00 AM
- Country of Publication:
- United States
- Language:
- English
Citation Formats
Michael, Peter, and Yoon, Hong-Jun. Survey of Image Denoising Methods for Medical Image Classification. United States: N. p., 2020.
Web. doi:10.1117/12.2549695.
Michael, Peter, & Yoon, Hong-Jun. Survey of Image Denoising Methods for Medical Image Classification. United States. https://doi.org/10.1117/12.2549695
Michael, Peter, and Yoon, Hong-Jun. 2020.
"Survey of Image Denoising Methods for Medical Image Classification". United States. https://doi.org/10.1117/12.2549695. https://www.osti.gov/servlets/purl/1648905.
@article{osti_1648905,
title = {Survey of Image Denoising Methods for Medical Image Classification},
author = {Michael, Peter and Yoon, Hong-Jun},
abstractNote = {Medical imaging devices, such as X-ray machines, inherently produce images that suffer from visual noise. Our objectives were to (i.) determine the effect of image denoising on a medical image classification task, and (ii.) determine if there exists a correlation between image denoising performance and medical image classification performance. We performed the medical image classification task on chest X-rays using the DenseNet-121 convolutional neural network (CNN) and used the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) metrics as the image denoising performance measures. We first found that different denoising methods can make a statistically significant difference in classification performance for select labels. We also found that denoising methods affect fine-tuned models more than randomly-initialized models and that fine-tuned models have significantly higher and more uniform performance than randomly-initialized models. Lastly, we found that there is no significant correlation between PSNR and SSIM values and classification performance for our task.},
doi = {10.1117/12.2549695},
url = {https://www.osti.gov/biblio/1648905},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Sun Mar 01 00:00:00 EST 2020},
month = {Sun Mar 01 00:00:00 EST 2020}
}