Publication: Iterative Hard Thresholding with Nonzero Index Initialization
0
0
Issued Date
2021
Resource Type
Language
eng
File Type
application/pdf
Other identifier(s)
2-s2.0-85119419643
Rights Holder(s)
Scopus
Bibliographic Citation
Proceedings - 1st International Conference on Smart Technologies Communication and Robotics, STCR 2021. Vol , No. (2021)
Suggested Citation
Tausiesakul B. Iterative Hard Thresholding with Nonzero Index Initialization. Proceedings - 1st International Conference on Smart Technologies Communication and Robotics, STCR 2021. Vol , No. (2021). doi:10.1109/STCR51658.2021.9589000 Retrieved from: https://hdl.handle.net/20.500.14740/8157
Author(s)
Abstract
Several methods for signal acquisition in compressed sensing were proposed in the past. Iterative hard thresholding (IHT) algorithm and its variants can be considered as a kind of those methods based on gradient descent. Unfortunately, when the objective function has many local minima, the steepest descent typically suffers from being misled into attaining those local minima. One way to facilitate the nonlinear search to be close to the global solution is the preparation for good initialization. In this work, a nonzero index according to signal sparsity is applied as the initial value for the IHT algorithms, instead of all zeros as in the former works. Numerical examples illustrate that the nonzero index initialization can provide lower normalized root-mean-square error of the acquired signal than the conventional all-zeros initialization, especially for numerous nonzero elements in the signal. © 2021 IEEE.