Skip to main content
main-content
Top

Hint

Swipe to navigate through the chapters of this book

2021 | OriginalPaper | Chapter

Segmenting Two-Dimensional Structures with Strided Tensor Networks

Authors : Raghavendra Selvan, Erik B. Dam, Jens Petersen

Published in: Information Processing in Medical Imaging

Publisher: Springer International Publishing

share
SHARE

Abstract

Tensor networks provide an efficient approximation of operations involving high dimensional tensors and have been extensively used in modelling quantum many-body systems. More recently, supervised learning has been attempted with tensor networks, primarily focused on tasks such as image classification. In this work, we propose a novel formulation of tensor networks for supervised image segmentation which allows them to operate on high resolution medical images. We use the matrix product state (MPS) tensor network on non-overlapping patches of a given input image to predict the segmentation mask by learning a pixel-wise linear classification rule in a high dimensional space. The proposed model is end-to-end trainable using backpropagation. It is implemented as a strided tensor network to reduce the parameter complexity. The performance of the proposed method is evaluated on two public medical imaging datasets and compared to relevant baselines. The evaluation shows that the strided tensor network yields competitive performance compared to CNN-based models while using fewer resources. Additionally, based on the experiments we discuss the feasibility of using fully linear models for segmentation tasks.(Source code: https://​github.​com/​raghavian/​strided-tenet)
Footnotes
1
First author of [22] noted their U-net work was cited more than once every hour in 2020. https://​bit.​ly/​unet2020.
 
2
Matrix product states are also known as Tensor Trains in literature.
 
3
Tensor product is the generalisation of matrix outer product to higher order tensors.
 
5
Tensor indices are dropped for brevity in the remainder of the manuscript.
 
8
These numbers are reported from [8] for their CNN2 model used for binary segmentation. Run time in Table 1 for CNN2 model could be lower with more recent hardware.
 
Literature
1.
go back to reference Anthony, L.F.W., Kanding, B., Selvan, R.: Carbontracker: tracking and predicting the carbon footprint of training deep learning models. In: ICML Workshop on Challenges in Deploying and monitoring Machine Learning Systems, July 2020. arXiv:​2007.​03051 Anthony, L.F.W., Kanding, B., Selvan, R.: Carbontracker: tracking and predicting the carbon footprint of training deep learning models. In: ICML Workshop on Challenges in Deploying and monitoring Machine Learning Systems, July 2020. arXiv:​2007.​03051
2.
go back to reference Bengua, J.A., Phien, H.N., Tuan, H.D., Do, M.N.: Matrix product state for feature extraction of higher-order tensors. arXiv preprint arXiv:​1503.​00516 (2015) Bengua, J.A., Phien, H.N., Tuan, H.D., Do, M.N.: Matrix product state for feature extraction of higher-order tensors. arXiv preprint arXiv:​1503.​00516 (2015)
3.
go back to reference Bridgeman, J.C., Chubb, C.T.: Hand-waving and interpretive dance: an introductory course on tensor networks. J. Phys. 50(22), 223001 (2017) Bridgeman, J.C., Chubb, C.T.: Hand-waving and interpretive dance: an introductory course on tensor networks. J. Phys. 50(22), 223001 (2017)
4.
go back to reference Cortes, C., Vapnik, V.: Support-vector networks. Mach. Learn. 20(3), 273–297 (1995) Cortes, C., Vapnik, V.: Support-vector networks. Mach. Learn. 20(3), 273–297 (1995)
6.
7.
go back to reference Jaeger, S., Candemir, S., Antani, S., Wáng, Y.X.J., Lu, P.X., Thoma, G.: Two public chest x-ray datasets for computer-aided screening of pulmonary diseases. Quant. Imaging Med. Surg. 4(6), 475 (2014) Jaeger, S., Candemir, S., Antani, S., Wáng, Y.X.J., Lu, P.X., Thoma, G.: Two public chest x-ray datasets for computer-aided screening of pulmonary diseases. Quant. Imaging Med. Surg. 4(6), 475 (2014)
8.
go back to reference Kumar, N., Verma, R., Sharma, S., Bhargava, S., Vahadane, A., Sethi, A.: A dataset and a technique for generalized nuclear segmentation for computational pathology. IEEE Trans. Med. Imaging 36(7), 1550–1560 (2017) CrossRef Kumar, N., Verma, R., Sharma, S., Bhargava, S., Vahadane, A., Sethi, A.: A dataset and a technique for generalized nuclear segmentation for computational pathology. IEEE Trans. Med. Imaging 36(7), 1550–1560 (2017) CrossRef
10.
go back to reference LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521(7553), 436–444 (2015) CrossRef LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521(7553), 436–444 (2015) CrossRef
11.
go back to reference MacKay, D.J.: Introduction to Gaussian processes. NATO ASI Series F Computer and Systems Sciences 168, 133–166 (1998) MATH MacKay, D.J.: Introduction to Gaussian processes. NATO ASI Series F Computer and Systems Sciences 168, 133–166 (1998) MATH
13.
go back to reference Novikov, A., Trofimov, M., Oseledets, I.: Exponential machines. Bull. Polish Acad. Sci. Tech. Sci. 66(6), 789–797 (2018) Novikov, A., Trofimov, M., Oseledets, I.: Exponential machines. Bull. Polish Acad. Sci. Tech. Sci. 66(6), 789–797 (2018)
14.
go back to reference Novikov, A., Izmailov, P., Khrulkov, V., Figurnov, M., Oseledets, I.V.: Tensor train decomposition on tensorflow (t3f). J. Mach. Learn. Res. 21(30), 1–7 (2020) Novikov, A., Izmailov, P., Khrulkov, V., Figurnov, M., Oseledets, I.V.: Tensor train decomposition on tensorflow (t3f). J. Mach. Learn. Res. 21(30), 1–7 (2020)
15.
go back to reference Novikov, A., Podoprikhin, D., Osokin, A., Vetrov, D.P.: Tensorizing neural networks. In: Advances in Neural Information Processing Systems, pp. 442–450 (2015) Novikov, A., Podoprikhin, D., Osokin, A., Vetrov, D.P.: Tensorizing neural networks. In: Advances in Neural Information Processing Systems, pp. 442–450 (2015)
16.
go back to reference Orús, R.: A practical introduction to tensor networks: matrix product states and projected entangled pair states. Ann. Phys. 349, 117–158 (2014) MathSciNetCrossRef Orús, R.: A practical introduction to tensor networks: matrix product states and projected entangled pair states. Ann. Phys. 349, 117–158 (2014) MathSciNetCrossRef
18.
19.
go back to reference Perez-Garcia, D., Verstraete, F., Wolf, M.M., Cirac, J.I.: Matrix product state representations. arXiv preprint quant-ph/0608197 (2006) Perez-Garcia, D., Verstraete, F., Wolf, M.M., Cirac, J.I.: Matrix product state representations. arXiv preprint quant-ph/0608197 (2006)
20.
go back to reference Poulin, D., Qarry, A., Somma, R., Verstraete, F.: Quantum simulation of time-dependent Hamiltonians and the convenient illusion of Hilbert space. Phys. Rev. Lett. 106(17), 170501 (2011) Poulin, D., Qarry, A., Somma, R., Verstraete, F.: Quantum simulation of time-dependent Hamiltonians and the convenient illusion of Hilbert space. Phys. Rev. Lett. 106(17), 170501 (2011)
21.
23.
go back to reference Selvan, R., Dam, E.B.: Tensor networks for medical image classification. In: International Conference on Medical Imaging with Deep Learning - Full Paper Track. Proceedings of Machine Learning Research, vol. 121, pp. 721–732. PMLR (06–08 Jul 2020) Selvan, R., Dam, E.B.: Tensor networks for medical image classification. In: International Conference on Medical Imaging with Deep Learning - Full Paper Track. Proceedings of Machine Learning Research, vol. 121, pp. 721–732. PMLR (06–08 Jul 2020)
24.
go back to reference Soares, J.V., Leandro, J.J., Cesar, R.M., Jelinek, H.F., Cree, M.J.: Retinal vessel segmentation using the 2-D Gabor wavelet and supervised classification. IEEE Trans. Med. Imaging 25(9), 1214–1222 (2006) CrossRef Soares, J.V., Leandro, J.J., Cesar, R.M., Jelinek, H.F., Cree, M.J.: Retinal vessel segmentation using the 2-D Gabor wavelet and supervised classification. IEEE Trans. Med. Imaging 25(9), 1214–1222 (2006) CrossRef
25.
go back to reference Stoudenmire, E., Schwab, D.J.: Supervised learning with tensor networks. In: Advances in Neural Information Processing Systems, pp. 4799–4807 (2016) Stoudenmire, E., Schwab, D.J.: Supervised learning with tensor networks. In: Advances in Neural Information Processing Systems, pp. 4799–4807 (2016)
26.
go back to reference Vermeer, K., Van der Schoot, J., Lemij, H., De Boer, J.: Automated segmentation by pixel classification of retinal layers in ophthalmic OCT images. Biomed. Opt. Express 2(6), 1743–1756 (2011) CrossRef Vermeer, K., Van der Schoot, J., Lemij, H., De Boer, J.: Automated segmentation by pixel classification of retinal layers in ophthalmic OCT images. Biomed. Opt. Express 2(6), 1743–1756 (2011) CrossRef
Metadata
Title
Segmenting Two-Dimensional Structures with Strided Tensor Networks
Authors
Raghavendra Selvan
Erik B. Dam
Jens Petersen
Copyright Year
2021
DOI
https://doi.org/10.1007/978-3-030-78191-0_31

Premium Partner