Automatic detection and segmentation of GGO tumor

Yuanjie Zheng, Xiaoyan(Ethel) Xiang, and Chandra Kambhamettu

Collaborators: Thomas Bauer and Karl Steiner

Ground-Glass Opacity(GGO) of a nodule can be observed in the High Resolution CT images by the hazy appearance of the image pixels, which are blurred and have less opacity compared to the usual "thick/opaque" nodule pixels. GGO plays an important role in categorizing the severity of a tumor. It is reported that traditional algorithms developed for segmentation of solid nodules are inaccurate when applied to GGO's. Therefore, many studies aimed at automatic and precise detection of GGO have been carried out in recent years. In this project, a working system for GGO lung nodule segmentation is developed, where a novel initialization methodology and Graph cut technique is used to guarantee high accuracy.

Fig 1. (a) GGO in CT image; (b) Labeled original CT image: foreground GGO pixels (red) and background (blue) other tissues; (c) Refined initializations by colorization technique; (d) Segmented results by Graph cut.

Fig 2. 3D reconstruction of a GGO from different views.

In the news

Related publications

  • Yuanjie Zheng, Chandra Kambhamettu, Thomas L. Bauer, Karl Steiner, Accurate Estimation of Pulmonary Nodule's Growth Rate in CT Images with Nonrigid Registration and Precise Nodule Detection and Segmentation, MMBIA 2009: Mathematical Methods in Biomedical Image Analysis, Miami, Florida, June 20, 2009.
  • Yuanjie Zheng, Chandra Kambhamettu, Thomas Bauer, Karl Steiner, Estimation of Ground-Glass Opacity Measurement in CT Lung Images, MICCAI 2008: International Conference on Medical Image Computing and Computer Assisted Intervention, New York City, USA, September, 2008.
  • Yuanjie Zheng, Karl Steiner, Thomas L. Bauer, Jingyi Yu, Dinggang Shen, Chandra Kambhamettu, Lung Nodule Growth Analysis from 3D CT Data with a Coupled Segmentation and Registration Framework, MMBIA 2007: IEEE Computer Society Workshop on Mathematical Methods in Biomedical Image Analysis, Rio de Janeiro, Brasil, Oct 14-15, 2007.