Bayesian Model for Liver Tumor Enhancement
Automatic liver lesion enhancement and detection has an essential role for the computer-aided diagnosis of liver tumor in CT volume data. This paper proposes a novel lesion enhancement strategy using Bayesian framework by combining the lesion probabilities based on an adaptive non-parametric model with the processed test volume and the constructed common non-lesion models with prepared liver database. Due to the large variation of different lesion tissues, it is difficult to obtain the common lesion prototypes from liver volumes, and thus this paper investigates a lesion-training-data free strategy by only constructing the healthy liver and vessel prototypes using local patches, which can be extracted from any slice of the test liver volume, and is also easy to prepare the common training non-lesion samples for all volumes. With the healthy liver and vessel prototypes from the test volume, an adaptive non-parametric model is constructed for estimating the lesion possibility, which is considered as the pixel likelihood to lesion region; the common model constructed using the pre-prepared liver database is used to estimate the pixel probability, which is defined as prior knowledge due to the used unvaried model. Finally, the posterior probabilities based on Bayesian theory are achieved for enhancing lesion regions. Experimental results validate that the proposed framework can not only detect almost small lesion regions but also greatly reduce falsely detect regions.
KeywordsLiver lesion Prototype Baysian framework Enhancement Adaptive non-parametric model CT volume
This research was supported in part by the Grant-in Aid for Scientific Research from the Japanese Ministry for Education, Science, Culture and Sports (MEXT) under the Grant No. 15H01130 and No. 15K00253, in part by the MEXT Support Program for the Strategic Research Foundation at Private Universities (2013-2017), and in part by the Recruitment Program of Global Experts HAIOU Program from Zhejiang, China.
- 1.National Cancer Center, Japan: Center for Cancer Control and Information Services. http://ganjoho.jp/public/statistics/pub/statistics01.html
- 3.Hame, Y., et al.: Image analysis for liver tumor ablation treatment planning, hands-on image processing 2009. Robotiker-TecnaliaGoogle Scholar
- 4.Mala, K., et al.: Neural network based texture analysis of liver tumor from computed tomography images. Int. J. Biomed. Sci. 2, 33–40 (2006)Google Scholar
- 5.Park, S.-J. et al.: Automatic Hepatic Tumor Segmentation Using Statistical Optimal Threshold, Computational Science-ICCS2005, vol. 3514, pp. 934–940. Springer, Berlin (2005)Google Scholar
- 6.Masuda, Y., et al.: Automatic liver tumor detection using EM/MPM algorithm and shape information. IEICE Technical Report, vol. 110, pp. 25–30 (2010)Google Scholar
- 7.Masuda, Y., et al.: Liver tumor detection In CT images by adaptive contrast enhancement and the EM/MPM algorithm. In: Proceedings of IEEE International Conference on Image Processing (ICIP2013), pp. 1453–1456 (2011)Google Scholar
- 9.Deng, X., Du, G.: Editorial: 3D segmentation in the clinic: a grand challenge II—liver tumor segmentation. http://grand-challenge2008.bigr.nl/proceedings/liver/articles.html
- 10.Foruzan, A.H., et al.: Segmentation of liver in low-contrast images using k-means clustering and geodesic active contour algorithms. IEICE Trans. E96-D, 798–807 (2013)Google Scholar
- 11.Dong, C., et al.: Simultaneous segmentation of multiple organs using random walks. J. Inf. Process. Soc. Jpn. 24, 320–329 (2016)Google Scholar