PROJECT TITLE :
A Fast Hadamard Transform for Signals With Sublinear Sparsity in the Transform Domain
During this paper, we tend to design a new iterative low-complexity algorithm for computing the Walsh–Hadamard rework (WHT) of an $N$ dimensional signal with a $K$ -sparse WHT. We tend to suppose that $N$ may be a power of 2 and $K=O(N^alpha)$ , scales sublinearly in $N$ for a few $alpha in (0,one)$ . Assuming a random support model for the nonzero transform-domain parts, our algorithm reconstructs the WHT of the signal with a sample complexity $O(K log _2(N/K))$ and a computational complexity $O(Klog _2(K) log _2(N/K))$ . Moreover, the algorithm succeeds with a high likelihood approaching one for massive dimension $N$ . Our approach is especially based mostly on the subsampling (aliasing) property of the WHT, where by a rigorously designed subsampling of the time-domain signal, a suitable aliasing pattern is induced within the transform domain. We have a tendency to treat the resulting aliasing patterns as parity-check constraints and represent them by a bipartite graph. We have a tendency to analyze the properties of the ensuing bipartite graphs and borrow ideas from codes outlined over sparse bipartite graphs to formulate the recovery of the nonzero spectral values as a peeling decoding algorithm for a particular sparse-graph code transmitted over a binary erasure channel. This allows us to use tools from coding theory (belief-propagation analysis) to characterize the- asymptotic performance of our algorithm within the very sparse ( $alpha in (zero, 1/3]$ ) and therefore the less sparse ( $alpha in (1/3, 1)$ ) regime. Comprehensive simulation results are provided to assess the empirical performance of the proposed algorithm.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here