Sparse Recovery with Very Sparse Compressed Counting
This work addresses compressed sensing for nonnegative data like images, offering an incremental improvement in sparsity with theoretical guarantees.
The paper tackles the problem of recovering nonnegative sparse signals using very sparse compressed counting, showing that with a maximally-skewed p-stable distribution and sparsified design matrix, the required measurements are M = K/(1-exp(-gK)) log N, leading to minor inflation in sample complexity compared to dense designs, such as M = 1.16K log N for g=2/K.
Compressed sensing (sparse signal recovery) often encounters nonnegative data (e.g., images). Recently we developed the methodology of using (dense) Compressed Counting for recovering nonnegative K-sparse signals. In this paper, we adopt very sparse Compressed Counting for nonnegative signal recovery. Our design matrix is sampled from a maximally-skewed p-stable distribution (0<p<1), and we sparsify the design matrix so that on average (1-g)-fraction of the entries become zero. The idea is related to very sparse stable random projections (Li et al 2006 and Li 2007), the prior work for estimating summary statistics of the data. In our theoretical analysis, we show that, when p->0, it suffices to use M= K/(1-exp(-gK) log N measurements, so that all coordinates can be recovered in one scan of the coordinates. If g = 1 (i.e., dense design), then M = K log N. If g= 1/K or 2/K (i.e., very sparse design), then M = 1.58K log N or M = 1.16K log N. This means the design matrix can be indeed very sparse at only a minor inflation of the sample complexity. Interestingly, as p->1, the required number of measurements is essentially M = 2.7K log N, provided g= 1/K. It turns out that this result is a general worst-case bound.