Compressing virtual pathology slides: Human & model observer evaluation

Elizabeth A. Krupinski, Jeffrey P. Johnson, Stacey Jaw, Anna R. Graham, Ronald S. Weinstein

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

Abstract

We aim to improve telepathology images for diagnoses using compression based on information about human visual system. Underlying goal is to demonstrate utility of a visual discrimination model (VDM) for predicting observer performance. 100 ROIs from breast biopsy virtual slides at 5 levels of compression (uncompressed, 8:1, 16:1, 32:1, 64:1, 128:1) were shown to 6 pathologists to determine benign vs malignant. There was a decrease in performance as a function of compression (F = 14.58, p< 0.0001). The visibility of compression artifacts in the test images was predicted using a VDM. JND metrics were computed for each image including mean, median, ≥90th percentiles, and maximum. For comparison PSNR and SSIM were also computed. Image distortion metrics were computed as a function of compression ratio and averaged across test images. All of the JND metrics were found to be highly correlated and differed primarily in magnitude. Both PSNR and SSIM decreased with bit rate, correctly reflecting a loss of image fidelity with increasing compression. The correlation of observer performance in the ROC experiment with image distortion metrics is shown in Figures 3 and 4. Observer performance (Az) was nearly constant up to a compression ratio of 32:1, then decreased significantly for 64:1 and 128:1 compression. The initial decline in Az occurred around a mean JND of 3, Minkowski JND of 4, and 99th percentile JND of 6.5. Virtual pathology may be compressible to relatively high levels before impacting diagnostic accuracy and the VDM accurately predicts human performance.

Original languageEnglish (US)
Title of host publicationMedical Imaging 2012
Subtitle of host publicationImage Perception, Observer Performance, and Technology Assessment
EditorsCraig K. Abbey, Claudia R. Mello-Thoms
PublisherSPIE
ISBN (Electronic)9780819489678
DOIs
StatePublished - Jan 1 2012
EventMedical Imaging 2012: Image Perception, Observer Performance, and Technology Assessment - San Diego, United States
Duration: Feb 8 2012Feb 9 2012

Publication series

NameProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume8318
ISSN (Print)1605-7422

Conference

ConferenceMedical Imaging 2012: Image Perception, Observer Performance, and Technology Assessment
CountryUnited States
CitySan Diego
Period2/8/122/9/12

Fingerprint

pathology
Pathology
chutes
compressing
Telepathology
visual discrimination
evaluation
Artifacts
Breast
Biopsy
compression ratio
Visibility
human performance
visibility
breast
Experiments
artifacts
Pathologists

Keywords

  • Compression
  • Model observer
  • Observer performance
  • Pathology virtual images

ASJC Scopus subject areas

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

Cite this

Krupinski, E. A., Johnson, J. P., Jaw, S., Graham, A. R., & Weinstein, R. S. (2012). Compressing virtual pathology slides: Human & model observer evaluation. In C. K. Abbey, & C. R. Mello-Thoms (Eds.), Medical Imaging 2012: Image Perception, Observer Performance, and Technology Assessment [83180Q] (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 8318). SPIE. https://doi.org/10.1117/12.911419

Compressing virtual pathology slides : Human & model observer evaluation. / Krupinski, Elizabeth A.; Johnson, Jeffrey P.; Jaw, Stacey; Graham, Anna R.; Weinstein, Ronald S.

Medical Imaging 2012: Image Perception, Observer Performance, and Technology Assessment. ed. / Craig K. Abbey; Claudia R. Mello-Thoms. SPIE, 2012. 83180Q (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 8318).

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

Krupinski, EA, Johnson, JP, Jaw, S, Graham, AR & Weinstein, RS 2012, Compressing virtual pathology slides: Human & model observer evaluation. in CK Abbey & CR Mello-Thoms (eds), Medical Imaging 2012: Image Perception, Observer Performance, and Technology Assessment., 83180Q, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 8318, SPIE, Medical Imaging 2012: Image Perception, Observer Performance, and Technology Assessment, San Diego, United States, 2/8/12. https://doi.org/10.1117/12.911419
Krupinski EA, Johnson JP, Jaw S, Graham AR, Weinstein RS. Compressing virtual pathology slides: Human & model observer evaluation. In Abbey CK, Mello-Thoms CR, editors, Medical Imaging 2012: Image Perception, Observer Performance, and Technology Assessment. SPIE. 2012. 83180Q. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE). https://doi.org/10.1117/12.911419
Krupinski, Elizabeth A. ; Johnson, Jeffrey P. ; Jaw, Stacey ; Graham, Anna R. ; Weinstein, Ronald S. / Compressing virtual pathology slides : Human & model observer evaluation. Medical Imaging 2012: Image Perception, Observer Performance, and Technology Assessment. editor / Craig K. Abbey ; Claudia R. Mello-Thoms. SPIE, 2012. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE).
@inproceedings{6dc54c9415384d1b8009be6e6976de02,
title = "Compressing virtual pathology slides: Human & model observer evaluation",
abstract = "We aim to improve telepathology images for diagnoses using compression based on information about human visual system. Underlying goal is to demonstrate utility of a visual discrimination model (VDM) for predicting observer performance. 100 ROIs from breast biopsy virtual slides at 5 levels of compression (uncompressed, 8:1, 16:1, 32:1, 64:1, 128:1) were shown to 6 pathologists to determine benign vs malignant. There was a decrease in performance as a function of compression (F = 14.58, p< 0.0001). The visibility of compression artifacts in the test images was predicted using a VDM. JND metrics were computed for each image including mean, median, ≥90th percentiles, and maximum. For comparison PSNR and SSIM were also computed. Image distortion metrics were computed as a function of compression ratio and averaged across test images. All of the JND metrics were found to be highly correlated and differed primarily in magnitude. Both PSNR and SSIM decreased with bit rate, correctly reflecting a loss of image fidelity with increasing compression. The correlation of observer performance in the ROC experiment with image distortion metrics is shown in Figures 3 and 4. Observer performance (Az) was nearly constant up to a compression ratio of 32:1, then decreased significantly for 64:1 and 128:1 compression. The initial decline in Az occurred around a mean JND of 3, Minkowski JND of 4, and 99th percentile JND of 6.5. Virtual pathology may be compressible to relatively high levels before impacting diagnostic accuracy and the VDM accurately predicts human performance.",
keywords = "Compression, Model observer, Observer performance, Pathology virtual images",
author = "Krupinski, {Elizabeth A.} and Johnson, {Jeffrey P.} and Stacey Jaw and Graham, {Anna R.} and Weinstein, {Ronald S.}",
year = "2012",
month = "1",
day = "1",
doi = "10.1117/12.911419",
language = "English (US)",
series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",
publisher = "SPIE",
editor = "Abbey, {Craig K.} and Mello-Thoms, {Claudia R.}",
booktitle = "Medical Imaging 2012",

}

TY - GEN

T1 - Compressing virtual pathology slides

T2 - Human & model observer evaluation

AU - Krupinski, Elizabeth A.

AU - Johnson, Jeffrey P.

AU - Jaw, Stacey

AU - Graham, Anna R.

AU - Weinstein, Ronald S.

PY - 2012/1/1

Y1 - 2012/1/1

N2 - We aim to improve telepathology images for diagnoses using compression based on information about human visual system. Underlying goal is to demonstrate utility of a visual discrimination model (VDM) for predicting observer performance. 100 ROIs from breast biopsy virtual slides at 5 levels of compression (uncompressed, 8:1, 16:1, 32:1, 64:1, 128:1) were shown to 6 pathologists to determine benign vs malignant. There was a decrease in performance as a function of compression (F = 14.58, p< 0.0001). The visibility of compression artifacts in the test images was predicted using a VDM. JND metrics were computed for each image including mean, median, ≥90th percentiles, and maximum. For comparison PSNR and SSIM were also computed. Image distortion metrics were computed as a function of compression ratio and averaged across test images. All of the JND metrics were found to be highly correlated and differed primarily in magnitude. Both PSNR and SSIM decreased with bit rate, correctly reflecting a loss of image fidelity with increasing compression. The correlation of observer performance in the ROC experiment with image distortion metrics is shown in Figures 3 and 4. Observer performance (Az) was nearly constant up to a compression ratio of 32:1, then decreased significantly for 64:1 and 128:1 compression. The initial decline in Az occurred around a mean JND of 3, Minkowski JND of 4, and 99th percentile JND of 6.5. Virtual pathology may be compressible to relatively high levels before impacting diagnostic accuracy and the VDM accurately predicts human performance.

AB - We aim to improve telepathology images for diagnoses using compression based on information about human visual system. Underlying goal is to demonstrate utility of a visual discrimination model (VDM) for predicting observer performance. 100 ROIs from breast biopsy virtual slides at 5 levels of compression (uncompressed, 8:1, 16:1, 32:1, 64:1, 128:1) were shown to 6 pathologists to determine benign vs malignant. There was a decrease in performance as a function of compression (F = 14.58, p< 0.0001). The visibility of compression artifacts in the test images was predicted using a VDM. JND metrics were computed for each image including mean, median, ≥90th percentiles, and maximum. For comparison PSNR and SSIM were also computed. Image distortion metrics were computed as a function of compression ratio and averaged across test images. All of the JND metrics were found to be highly correlated and differed primarily in magnitude. Both PSNR and SSIM decreased with bit rate, correctly reflecting a loss of image fidelity with increasing compression. The correlation of observer performance in the ROC experiment with image distortion metrics is shown in Figures 3 and 4. Observer performance (Az) was nearly constant up to a compression ratio of 32:1, then decreased significantly for 64:1 and 128:1 compression. The initial decline in Az occurred around a mean JND of 3, Minkowski JND of 4, and 99th percentile JND of 6.5. Virtual pathology may be compressible to relatively high levels before impacting diagnostic accuracy and the VDM accurately predicts human performance.

KW - Compression

KW - Model observer

KW - Observer performance

KW - Pathology virtual images

UR - http://www.scopus.com/inward/record.url?scp=85076808527&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85076808527&partnerID=8YFLogxK

U2 - 10.1117/12.911419

DO - 10.1117/12.911419

M3 - Conference contribution

AN - SCOPUS:85076808527

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

BT - Medical Imaging 2012

A2 - Abbey, Craig K.

A2 - Mello-Thoms, Claudia R.

PB - SPIE

ER -