Segmentation in high dimensional space, e.g. 4D, often requires decomposition of the space and sequential data process, for instance space followed by time. In, the authors presented a deformable model that can be generalized into arbitrary dimensions. However, its direct implementation is computationally prohibitive. The more efficient method proposed by the same authors has significant overhead on computer memory, which is not desirable for high dimensional data processing. In this work, we propose a novel approach to formulate the computation to achieve memory efficiency, as well as improving computational efficiency. Numerical studies on synthetic data and preliminary results on real world data suggest that the proposed method has a great potential in biomedical applications where data is often inherently high dimensional.
展开▼
