Solitary Pulmonary Nodule Detection on Thoracic CT Images through Object Continuity Analyses

Abstract:

Lung cancer has been a leading cause of death in the world, and it is known that prompt diagnosis and treatment may be the only chance for curing the cancer. Early lung cancer often presents as a solitary pulmonary nodule (SPN) and the timely detection of it is critical to save life from cancer death. In this paper, we present an effective method to detect SPNs on thoracic CT images through object continuity analyses. First, a lung region is segmented from other chest organs using morphological operations and thresholding techniques, and an initial set of candidate SPNs are identified. To represent the SPN, we define the rotation-invariant bounding rectangle (riBR) that tightly encloses an object. The subsequent processing is based on the riBR instead of an object itself to avoid the processing overhead. Next, non-nodule objects are pruned using geometric features and the object continuity analyses on a series of CT slice images. Through the analyses, cylinder-shaped non-nodule objects such as blood vessels and bronchia are eliminated and a final set of candidate SPNs is obtained. An experimental result shows that the proposed method works effectively in detecting SPNs. The application context addressed in this study is the pulmonary nodule detection but other application areas also can benefit.

• Periodical: Applied Mechanics and Materials (Volumes 284-287)
• Pages: 1681-1685
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.284-287.1681
• Online since: January 2013
• Authors: Seok Lyong Lee, Du Hyung Cho
• Keywords: Nodule Detection, Object Continuity Analysis, Object Recognition, Solitary Pulmonary Nodule, Thoracic CT Image

References

[1] S. Sone, S. Takashima, F. Li, Z. Yang, T. Honda, Y. Maruyama, M. Hasegawa, T. Yamanda, K. Kubo, K. Hanamura, K. Asakura, The Lancet, 351, 1242 (1998).

DOI: 10.1016/s0140-6736(97)08229-9

[2] T. Nawa, T. Nakagawa, S. Kusano, Y. Kawasaki, Y. Sugawara, H. Nakata, Chest, 122, 15 (2002).

DOI: 10.1378/chest.122.1.15

[3] J. M. Stephen, A. P. Maxine, W. R. Michael, Current Medical Diagnosis and Treatment 2012, McGraw-Hill, (2012).

[4] Y. Liu, Z. Xing, C. Deng, P. Li, M. Guo, Automatically detecting lung nodules based on shape descriptor and semi-supervised learning, Proceedings of International Conference on the Computer Application and System Modeling, (2010).

DOI: 10.1109/iccasm.2010.5619447

[5] X. Ye, X. Lin, J. Dehmeshki, G. Slabaugh, G. Beddoe, IEEE Tr. Med. Imaging, 56, 1810 (2009).

[6] A. Farag, J. Graham, S. Elshazly, A. Farag, Statistical modeling of the lung nodules in low dose computed tomography scans of the chest, Proceedings of the 17th IEEE International Conference on Image Processing, (2010).

DOI: 10.1109/icip.2010.5651832

[7] Y. Lee, T. Hara, H. Fujita, S. Itoh, T. Ishigaki, IEEE Tr. Med. Imaging, 20, 595 (2001).

[8] M. Antonelli, G. Frosini, B. Lazzerini, F. Marcelloni, World Academy of Sci., Eng. and Tech., 1, 128 (2005).

[9] G. Agam, S. G. Armato III, C. Wu, IEEE Tr. Med. Imaging, 24, 486 (2005).

[10] J. Dehmeshki, H. Amin, M. Valdivieso, X. Ye, IEEE Tr. Med. Imaging, 27, 467 (2008).

[11] N. A. Otsu, IEEE Tr. System Man and Cybernetics, 9, 62 (1979).