Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2012;15(Pt 3):459-67.
doi: 10.1007/978-3-642-33454-2_57.

Co-segmentation of functional and anatomical images

Ulas Bagci  1 , Jayaram K Udupa, Jianhua Yao, Daniel J Mollura
Affiliations

Co-segmentation of functional and anatomical images

Ulas Bagci et al. Med Image Comput Comput Assist Interv. 2012.

Abstract

This paper presents a novel method for segmenting functional and anatomical structures simultaneously. The proposed method unifies domains of anatomical and functional images (PET-CT), represents them in a product lattice, and performs simultaneous delineation of regions based on a random walk image segmentation. In addition, we propose a simple yet efficient object/background seed localization method, where background and foreground object cues are automatically obtained from PET images and propagated onto the corresponding anatomical images (CT). In our experiments, abnormal anatomies on PET-CT images from human subjects are segmented synergistically by the proposed fully automatic co-segmentation method with high precision (mean DSC of 91.44%) in seconds (avg. 40 seconds).

Keywords: Joint Segmentation; Object Detection; PET-CT; Random Walk.

PubMed Disclaimer

Figures

Fig. 1
Fig. 1
Even though intensity profiles in CT images (b) show similar characteristics, SUV from PET images (a) might further characterize the nodules. Radioactivity uptake regions are shown in small nodules in (a and c)(and details of fused image (c) are shown in d and e). The concepts of interesting uptake region (IUR) detection and background/foreground seed localization are sketched in (f–i).
Fig. 2
Fig. 2
(a) Automatically located background (red) and foreground (blue) seeds. (b) Additional background seeds are obtained by connecting initial background seeds using b-splines. (c) Ground truth (black) and random walk segmentations using only a limited number of background seeds. (d) The effect of the proposed seed localization method (with additional background seeds) in avoiding possible leakages.
Fig. 3
Fig. 3
Two different segmentation examples of uptake regions are shown in each column. First column: co-segmentation (blue) and ground truth (black) are overlaid. Second column: ground truth (black) and segmentation from PET only (yellow). Third column: ground truth (black) and segmentation from CT only (green). Fourth column: all segmentations and ground truth are overlaid together.
Fig. 4
Fig. 4
Mean DSCs and HDs are enlisted. DSC ratios: PET Only: 83.23 ∓ 1.87, CT Only: 87.88 ∓ 2.04, Han et al.: 89.34 ∓ 1.95, PET-CT cosegm.: 91.44 ∓ 1.71. HDs ratios: PET Only: 5.25 ∓ 0.53, CT Only: 4.82 ∓ 0.38, Han et al: 4.65 ∓ 0.73, PET-CT cosegm.:4.47 ∓ 0.54.
See this image and copyright information in PMC

References

    1. Judenhofer MS, et al. Simultaneous PET-MRI: a new approach for functional and morphological imaging. Nature Medicine. 2008;14:459–465. - PubMed
    1. Han D, et al. Globally optimal tumor segmentation in PET-CT images: a graph-based co-segmentation method. Inf Process Med Imaging. 2011:245–256. - PMC - PubMed
    1. Montgomery DW, et al. Fully automated segmentation of oncological PET volumes using a combined multiscale and statistical model. Medical Physics. 2007;34(2):722–736. - PubMed
    1. Jentzen W, et al. Segmentation of PET Volumes by Iterative Image Thresholding. J Nucl Med. 2007;48:108–114. - PubMed
    1. Grady L. Random Walks for Image Segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2006;28(11):1768–1783. - PubMed

MeSH terms