Learning task-specific and shared representations in medical imaging

TY - GEN

T1 - Learning task-specific and shared representations in medical imaging

AU - Bragman, Felix J.S.

AU - Tanno, Ryutaro

AU - Ourselin, Sebastien

AU - Alexander, Daniel C.

AU - Cardoso, M. Jorge

PY - 2019

Y1 - 2019

N2 - The performance of multi-task learning hinges on the design of feature sharing between tasks; a process which is combinatorial in the network depth and task count. Hand-crafting an architecture based on human intuitions of task relationships is therefore suboptimal. In this paper, we present a probabilistic approach to learning task-specific and shared representations in Convolutional Neural Networks (CNNs) for multi-task learning of semantic tasks. We introduce Stochastic Filter Groups; which is a mechanism that groups convolutional kernels into task-specific and shared groups to learn an optimal kernel allocation. They facilitate learning optimal shared and task specific representations. We employ variational inference to learn the posterior distribution over the possible grouping of kernels and CNN weights. Experiments on MRI-based prostate radiotherapy organ segmentation and CT synthesis demonstrate that the proposed method learns optimal task allocations that are inline with human-optimised networks whilst improving performance over competing baselines.

AB - The performance of multi-task learning hinges on the design of feature sharing between tasks; a process which is combinatorial in the network depth and task count. Hand-crafting an architecture based on human intuitions of task relationships is therefore suboptimal. In this paper, we present a probabilistic approach to learning task-specific and shared representations in Convolutional Neural Networks (CNNs) for multi-task learning of semantic tasks. We introduce Stochastic Filter Groups; which is a mechanism that groups convolutional kernels into task-specific and shared groups to learn an optimal kernel allocation. They facilitate learning optimal shared and task specific representations. We employ variational inference to learn the posterior distribution over the possible grouping of kernels and CNN weights. Experiments on MRI-based prostate radiotherapy organ segmentation and CT synthesis demonstrate that the proposed method learns optimal task allocations that are inline with human-optimised networks whilst improving performance over competing baselines.

UR - http://www.scopus.com/inward/record.url?scp=85075673495&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-32251-9_41

DO - 10.1007/978-3-030-32251-9_41

M3 - Conference contribution

AN - SCOPUS:85075673495

SN - 9783030322502

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 374

EP - 383

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2019 - 22nd International Conference, Proceedings

A2 - Shen, Dinggang

A2 - Yap, Pew-Thian

A2 - Liu, Tianming

A2 - Peters, Terry M.

A2 - Khan, Ali

A2 - Staib, Lawrence H.

A2 - Essert, Caroline

A2 - Zhou, Sean

PB - SPRINGER

T2 - 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2019

Y2 - 13 October 2019 through 17 October 2019

ER -