Real Sparse Fast DCT for Vectors with Short Support
For signal processing applications requiring fast reconstruction of sparse signals from DCT measurements, this provides a more efficient alternative to standard IDCT when the support size is small.
This paper presents a deterministic algorithm for the inverse discrete cosine transform (IDCT-II) that reconstructs a vector with short support from its DCT coefficients, achieving runtime O(M log M + m log(N/M)) and sampling O(M + m log(N/M)).
In this paper we present a new fast and deterministic algorithm for the inverse discrete cosine transform of type II for reconstructing the input vector $\mathbf x\in\mathbb R^N$, $N=2^J$, with short support of length $m$ from its discrete cosine transform $\mathbf x^{\widehat{\mathrm{II}}}=C^{\mathrm{II}}_N\mathbf x$ if an upper bound $M\geq m$ is known. The resulting algorithm only uses real arithmetic, has a runtime of $\mathcal{O}\left(M\log M+m\log_2\frac{N}{M}\right)$ and requires $\mathcal{O}\left(M+m\log_2\frac{N}{M}\right)$ samples of $\mathbf x^{\widehat{\mathrm{II}}}$. For $m,M\rightarrow N$ the runtime and sampling requirements approach those of a regular IDCT-II for vectors with full support. The algorithm presented hereafter does not employ inverse FFT algorithms to recover $\mathbf x$.