Hankel Matrix Nuclear Norm Regularized Tensor Completion for N-dimensional Exponential Signals

Signals are usually modeled as a superposition of exponential functions in spectroscopy of chemistry, biology and medical imaging. However, for fast data acquisition or other inevitable reasons, only a small amount of samples may be acquired. How to recover the full signal is then of great interest. Existing approaches can not efficiently recover N-dimensional exponential signals with N>=3. This paper studies the problem of recovering N-dimensional (particularly $N\geq 3$) exponential signals from partial observations, and we formulate this problem as a low-rank tensor completion problem with exponential factors. The full signal is reconstructed by simultaneously exploiting the CANDECOMP/PARAFAC (CP) tensor decomposition and the exponential structure of the associated factors, of which the latter is promoted by minimizing an objective function involving the nuclear norm of Hankel matrices. Experimental results on simulated and real magnetic resonance spectroscopy data show that the proposed approach can successfully recover full signals from very limited samples and is robust to the estimated tensor rank.