Feasibility study of deep convolutional generative adversarial networks to generate mammography images

Gihun Kim, Hyunjung Shim, Jongduk Baek

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

We conducted a feasibility study to generate mammography images using a deep convolutional generative adversarial network (DCGAN), which directly produces realistic images without 3-D model passing through any complex rendering algorithm, such as ray tracing. We trained DCGAN with breast 2D mammography images, which were generated from anatomical noise. The generated X-ray mammography images were successful in that the image preserves reasonable quality and retains the visual patterns similar to training images. Especially, generated images share the distinctive structure of training images. For the quantitative evaluation, we used the mean and variance of beta values of generated images and observed that they are very similar to those of training images. Although the general distribution of generated images matches well with those of training images, there are several limitations of the DCGAN. First, checkboard pattern like artifacts are found in generated images, which is a well-known issue of deconvolution algorithm. Moreover, training GAN is often unstable so to require manual fine-tunes. To overcome such limitations, we plan to extend our idea to conditional GAN approach for improving training stability, and employ an auto-encoder for handling artifacts. To validate our idea on real data, we will apply clinical images to train the network. We believe that our framework can be easily extended to generate other medical images.

Original languageEnglish
Title of host publicationMedical Imaging 2018
Subtitle of host publicationImage Perception, Observer Performance, and Technology Assessment
EditorsFrank W. Samuelson, Robert M. Nishikawa
PublisherSPIE
ISBN (Electronic)9781510616431
DOIs
Publication statusPublished - 2018 Jan 1
EventMedical Imaging 2018: Image Perception, Observer Performance, and Technology Assessment - Houston, United States
Duration: 2018 Feb 112018 Feb 12

Publication series

NameProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume10577
ISSN (Print)1605-7422

Conference

ConferenceMedical Imaging 2018: Image Perception, Observer Performance, and Technology Assessment
CountryUnited States
CityHouston
Period18/2/1118/2/12

Fingerprint

Mammography
Feasibility Studies
Artifacts
Three-Dimensional Imaging
Ray tracing
Deconvolution
education
Noise
Breast
X-Rays
X rays
artifacts
coders
ray tracing
breast

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Biomaterials
  • Atomic and Molecular Physics, and Optics
  • Radiology Nuclear Medicine and imaging

Cite this

Kim, G., Shim, H., & Baek, J. (2018). Feasibility study of deep convolutional generative adversarial networks to generate mammography images. In F. W. Samuelson, & R. M. Nishikawa (Eds.), Medical Imaging 2018: Image Perception, Observer Performance, and Technology Assessment [105771C] (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 10577). SPIE. https://doi.org/10.1117/12.2293046
Kim, Gihun ; Shim, Hyunjung ; Baek, Jongduk. / Feasibility study of deep convolutional generative adversarial networks to generate mammography images. Medical Imaging 2018: Image Perception, Observer Performance, and Technology Assessment. editor / Frank W. Samuelson ; Robert M. Nishikawa. SPIE, 2018. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE).
@inproceedings{e213449c4b594432aea0b5885ebcb871,
title = "Feasibility study of deep convolutional generative adversarial networks to generate mammography images",
abstract = "We conducted a feasibility study to generate mammography images using a deep convolutional generative adversarial network (DCGAN), which directly produces realistic images without 3-D model passing through any complex rendering algorithm, such as ray tracing. We trained DCGAN with breast 2D mammography images, which were generated from anatomical noise. The generated X-ray mammography images were successful in that the image preserves reasonable quality and retains the visual patterns similar to training images. Especially, generated images share the distinctive structure of training images. For the quantitative evaluation, we used the mean and variance of beta values of generated images and observed that they are very similar to those of training images. Although the general distribution of generated images matches well with those of training images, there are several limitations of the DCGAN. First, checkboard pattern like artifacts are found in generated images, which is a well-known issue of deconvolution algorithm. Moreover, training GAN is often unstable so to require manual fine-tunes. To overcome such limitations, we plan to extend our idea to conditional GAN approach for improving training stability, and employ an auto-encoder for handling artifacts. To validate our idea on real data, we will apply clinical images to train the network. We believe that our framework can be easily extended to generate other medical images.",
author = "Gihun Kim and Hyunjung Shim and Jongduk Baek",
year = "2018",
month = "1",
day = "1",
doi = "10.1117/12.2293046",
language = "English",
series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",
publisher = "SPIE",
editor = "Samuelson, {Frank W.} and Nishikawa, {Robert M.}",
booktitle = "Medical Imaging 2018",
address = "United States",

}

Kim, G, Shim, H & Baek, J 2018, Feasibility study of deep convolutional generative adversarial networks to generate mammography images. in FW Samuelson & RM Nishikawa (eds), Medical Imaging 2018: Image Perception, Observer Performance, and Technology Assessment., 105771C, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 10577, SPIE, Medical Imaging 2018: Image Perception, Observer Performance, and Technology Assessment, Houston, United States, 18/2/11. https://doi.org/10.1117/12.2293046

Feasibility study of deep convolutional generative adversarial networks to generate mammography images. / Kim, Gihun; Shim, Hyunjung; Baek, Jongduk.

Medical Imaging 2018: Image Perception, Observer Performance, and Technology Assessment. ed. / Frank W. Samuelson; Robert M. Nishikawa. SPIE, 2018. 105771C (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 10577).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Feasibility study of deep convolutional generative adversarial networks to generate mammography images

AU - Kim, Gihun

AU - Shim, Hyunjung

AU - Baek, Jongduk

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We conducted a feasibility study to generate mammography images using a deep convolutional generative adversarial network (DCGAN), which directly produces realistic images without 3-D model passing through any complex rendering algorithm, such as ray tracing. We trained DCGAN with breast 2D mammography images, which were generated from anatomical noise. The generated X-ray mammography images were successful in that the image preserves reasonable quality and retains the visual patterns similar to training images. Especially, generated images share the distinctive structure of training images. For the quantitative evaluation, we used the mean and variance of beta values of generated images and observed that they are very similar to those of training images. Although the general distribution of generated images matches well with those of training images, there are several limitations of the DCGAN. First, checkboard pattern like artifacts are found in generated images, which is a well-known issue of deconvolution algorithm. Moreover, training GAN is often unstable so to require manual fine-tunes. To overcome such limitations, we plan to extend our idea to conditional GAN approach for improving training stability, and employ an auto-encoder for handling artifacts. To validate our idea on real data, we will apply clinical images to train the network. We believe that our framework can be easily extended to generate other medical images.

AB - We conducted a feasibility study to generate mammography images using a deep convolutional generative adversarial network (DCGAN), which directly produces realistic images without 3-D model passing through any complex rendering algorithm, such as ray tracing. We trained DCGAN with breast 2D mammography images, which were generated from anatomical noise. The generated X-ray mammography images were successful in that the image preserves reasonable quality and retains the visual patterns similar to training images. Especially, generated images share the distinctive structure of training images. For the quantitative evaluation, we used the mean and variance of beta values of generated images and observed that they are very similar to those of training images. Although the general distribution of generated images matches well with those of training images, there are several limitations of the DCGAN. First, checkboard pattern like artifacts are found in generated images, which is a well-known issue of deconvolution algorithm. Moreover, training GAN is often unstable so to require manual fine-tunes. To overcome such limitations, we plan to extend our idea to conditional GAN approach for improving training stability, and employ an auto-encoder for handling artifacts. To validate our idea on real data, we will apply clinical images to train the network. We believe that our framework can be easily extended to generate other medical images.

UR - http://www.scopus.com/inward/record.url?scp=85047855545&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047855545&partnerID=8YFLogxK

U2 - 10.1117/12.2293046

DO - 10.1117/12.2293046

M3 - Conference contribution

AN - SCOPUS:85047855545

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2018

A2 - Samuelson, Frank W.

A2 - Nishikawa, Robert M.

PB - SPIE

ER -

Kim G, Shim H, Baek J. Feasibility study of deep convolutional generative adversarial networks to generate mammography images. In Samuelson FW, Nishikawa RM, editors, Medical Imaging 2018: Image Perception, Observer Performance, and Technology Assessment. SPIE. 2018. 105771C. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE). https://doi.org/10.1117/12.2293046