Tensor-Based Subspace Learning for Classification of Focal Liver Lesions in Multi-phase CT Images
Medical images play an important role in clinics. Machine learning has been widely used in the fields of computer vision and pattern recognition and computer-aided diagnosis with medical images become an active research topic. Efficient representation of the medical images or effective extraction of discriminative features from CT images is one of crucial steps in computer-aided diagnosis. Principal component analysis (PCA) is a subspace learning method and is widely used for efficient representation of data. The limitation of PCA is that a multi-dimensional data (e.g. an image or a video image) should be unfolded into a vector resulting in loss of spatial and spatial-temporal relationship of the data. In this paper, we proposed an efficient representation of multi-phase CT images based on a tensor-based subspace learning method known as generalized N-dimensional principal component analysis (GND-PCA). In the proposed method, the multi-phase CT image is treated a tensor without vector-unfolding for subspace learning. The core tensor obtained by GND-PCA is used as temporal and spatial features for focal liver lesion classification. Experiments show that in the case of fewer samples, GND-PCA achieved better results than conventional PCA and 2D-PCA, which is an extension of PCA.
KeywordsGeneralized N-dimensional principal component analysis (GND-PCA) Multi-phase medical image Focal liver lesion (FLL)
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. 18H03267, and No. 18H04747, and in part by Zhejiang Lab Program under the Grant No. 2018DG0ZX01.
- 6.Xu, Y., et al.: Bag of temporal co-occurrence words for retrieval of focal liver lesions using 3D multiphase contrast-enhanced CT images. In: Proceedings of 23rd International Conference on Pattern Recognition, ICPR 2016, pp. 2283–2288 (2016)Google Scholar
- 7.Wang, J., et al.: Sparse codebook model of local structures for retrieval of focal liver lesions using multiphase medical images. Int. J. Biomed. Imaging 2017, 13 (2017)Google Scholar
- 9.Frid-Adar, M., Diamant, I., Klang, E., Amitai, M., Goldberger, J., Greenspan, H.: Modeling the intra-class variability for liver lesion detection using a multi-class patch-based CNN. In: Wu, G., Munsell, B.C., Zhan, Y., Bai, W., Sanroma, G., Coupé, P. (eds.) Patch-MI 2017. LNCS, vol. 10530, pp. 129–137. Springer, Cham (2017). https://doi.org/10.1007/978-3-31967434-6_15
- 11.Liang, D., et al.: Residual convolutional neural networks with global and local pathways for classification of focal liver lesions. In: Geng, X., Kang, B.H. (eds.) PRICAI 2018: Trends in Artificial Intelligence, PRICAI 2018, Nanjin, China, 28–31 August 2018. Lecture Notes in Artificial Intelligence, vol. 11012, pp. 617–628. Springer (2018)Google Scholar
- 12.Liang, D., et al.: Combining Convolutional and recurrent neural networks for classification of focal liver lesions in multi-phase CT images. In: Frangi, A., Schnabel, J., Davatzikos, C., Alberola-López, C., Fichtinger, G. (eds.) Medical Image Computing and Computer Assisted Intervention – MICCAI 2018. Lecture Notes in Computer Science, LNCS, vol. 11071, pp. 666–675, Springer (2018)Google Scholar