Xiangrong Zhou PhD, Abstract Co-Author: Nothing to Disclose

Takaaki Ito, Abstract Co-Author: Nothing to Disclose

Takeshi Hara PhD, Abstract Co-Author: Nothing to Disclose

Ryujiro Yokoyama, Abstract Co-Author: Nothing to Disclose

Huayue Chen, Abstract Co-Author: Nothing to Disclose

Masayuki Kanematsu MD, Abstract Co-Author: Nothing to Disclose

Hiroshi Fujita PhD, Presenter: Nothing to Disclose

Background removal is a useful tool that is implemented in Microsoft Office 2010, and can be used to efficiently separate the target (foreground) from the background on a 2D picture. The algorithm of this tool is called “Grabcut” that accomplishes a fast and automatic background removal after marquee selection by indicating a 2D box region. The question whether such kind of background removal techniques work well for organ segmentations on clinical CT image is what we want to answer here.

A dataset consisting of more than 100 volumetric torso CT scans was used for testing. These CT images were generated from 3 kinds of CT scanners from GE, Toshiba, and Philips by using different scan protocols for clinical diagnosis. The spatial resolution of the CT images was varied from 0.62 mm to 2 mm. Nine major organs and tissues including heart, liver, spleen, left-kidney, right-kidney, bladder, gallbladder, left-psoas-major-muscle, and right-psoas-major- muscle were selected as segmentation targets. The Grabcut algorithm was expanded to 3D by our group to adapt to CT images. The Jaccard similarity coefficient (JSC) between the segmentation result and doctors's manual sketch was used as the accuracy measure.

Our evaluation showed that the average JSC values for these targets were in the range of 60 % to 70 %, and can be improved to 90 % by combining with an probabilistic atlas approach. Typical computing time for organ segmentation on a torso CT scan using a general-purpose computer equipped with an Intel Core2Duo 2.23-GHz CPU was less than 1 minute.

Background removal technique is helpful for the organ segmentations on CT images, but it was not as powerful as it is used on nature pictures and need to be improved.

http://abstract.rsna.org/uploads/2014/14017524/14017524_g5c8.jpg

Zhou, X, Ito, T, Hara, T, Yokoyama, R, Chen, H, Kanematsu, M, Fujita, H, Does the Background Removal Tool used on Nature Pictures Work Well for Organ Segmentations on 3D CT Images?.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14017524.html