News

Text Size

Media Briefing: New Technology Could Aid in Interpretation of Medical Imagery
10.14.10
 
NASA hosted a media teleconference at 11 a.m. ET on Thursday, Oct. 14, 2010, to discuss NASA software that has been incorporated into a new medical imaging technology that one day could aid in the interpretation of mammograms, ultrasounds, and other medical imagery.

NASA hosted this event in conjunction with Breast Cancer Awareness Month.

› NASA press release
› NASA feature story

› Audio recording of briefing (mp3)
› Transcript of briefing (pdf)




Briefing Speakers


  • Nona Cheeks, chief of the Innovative Partnerships Program Office, NASA's Goddard Space Flight Center, Greenbelt, Md.
  • James C. Tilton, Ph.D., computer engineer, NASA's Goddard Space Flight Center, Greenbelt, Md.
  • Fitz Walker, president and CEO of Bartron Medical Imaging
  • Dr. Molly Brewer, professor with the Division of Gynecologic Oncology, University of Connecticut Health Center




Images and Multimedia in Support of the News Conference


Presenter: Nona Cheeks, chief of the Innovative Partnerships Program Office, NASA's Goddard Space Flight Center

No visuals.



Presenter: James C. Tilton, Ph.D., computer engineer, NASA's Goddard Space Flight Center

true color depiction of a small portion of a Landsat Thematic Mapper (TM) scene This is a true-color depiction of a small portion of a Landsat Thematic Mapper (TM) scene of lakes in northern Wisconsin.
› Larger image
depiction of a segmentation/classification result from HSEG for the same Landsat TM scene This image is a depiction of a segmentation/classification result from HSEG for the same Landsat TM scene. Lighter toned lakes are colored light blue in the segmentation map, darker toned lakes are colored dark blue, the green colored areas are mainly vegetation, the yellow colored areas are probably cleared soil, roads and evidence of other human activity, and the brown areas are mixtures of ground cover.
› Larger image

depiction of a coarser resolution segmentation/classification result from HSEG for the same Landsat TM scene This image is a depiction of a coarser resolution segmentation/classification result from HSEG for the same Landsat TM scene. Here the scene is separated into just two classes: Lakes (including rivers) and Other.
› Larger image
unprocessed cell image An unprocessed cell image from the “Exploratory Image Set from the NIST CompBio Project Subcellular Feature Extraction Challenge Problem,” collected by the Cell Systems Science Group, National Institute of Standards and Technology (NIST) and provided by Alden Dima (NIST).
› Larger image
cell image with enhancement provided by the histogram equalization enhancement procedure The enhancement provided by the histogram equalization enhancement procedure. The histogram equalization enhancement procedure attempts to enhance an image through creating a remapping of the image data values that flattens the image’s histogram. A histogram is a plot of the frequency of occurrence of data values in an image.
› Larger image

cell image with enhancement provided by processing with HSEG The enhancement provided by processing with HSEG.
› Larger image
histogram of the original cell image Histogram of the original cell image.
› Larger image
histogram of the histogram equalized enhanced image Histogram of the histogram-equalized enhanced image.
› Larger image
histogram of the HSEG enhanced image Histogram of the HSEG enhanced image.
› Larger image
A histogram is a plot of the frequency of occurrence of data values in an image. In these plots you can see how the histogram equalization procedure spreads out or “flattens” the histogram of the image as much as possible. The HSEG enhancement procedure also tends to spread out the image histogram, but performs this spreading selectively utilizing information from the segmentation process: it spreads out the data values occurring in the cell more that it spreads out the data values occurring in the background.



Presenter: Fitz Walker, president and CEO of Bartron Medical Imaging

mammogram before MED-SEG processing Original mammogram before MED-SEG processing. Credit: Bartron Medical Imaging
› Larger image

mammogram after MED-SEG processing Mammogram, with region of interest (white) labeled, after MED-SEG processing. Credit: Bartron Medical Imaging
› Similar larger image




Presenter: Dr. Molly Brewer, professor with the Division of Gynecologic Oncology, University of Connecticut Health Center

No visuals.



Supplemental Images


true color rendition of an Ikonos satellite image from over Patterson Park in Baltimore, MD A true-color rendition of a 384x384 pixel portion of an Ikonos satellite image from over Patterson Park in Baltimore, Md. Credit:
› Larger image
the region average image from an HSEG segmentation The region average image from an HSEG segmentation. Credit:
› Larger image
the region average image from a HSWO segmentation The region average image from a Hierarchical Step-Wise Optimization (HSWO) segmentation with the same maximum merge threshold used for the HSEG segmentation in the middle image. HSWO is similar to HSEG but without the classification step. Comparing these two results shows the effect of the HSEG region classification step on the segmentation result. Credit:
› Larger image



Related Video