Export
Pancham Shukla
· 1 previous affiliation
- Imperial College LondonSince 2021
- Lecturer · Since 2021
- PhD student · 2003–2006
Research Overview
Pancham Shukla focuses on signal processing, finite rate of innovation sampling, and multidimensional signal reconstruction. The available evidence captures a concentrated period of research developing novel sampling schemes for nonband-limited signals, particularly two-dimensional and multidimensional signals. The most recent work in the corpus addresses generalized frameworks for perfectly reconstructing these signals using specialized kernels.
Research Trajectory
The available evidence for Pancham Shukla’s research trajectory is highly concentrated, spanning the years 2005 to 2007. During this period, Shukla led a focused research agenda on the sampling and reconstruction of signals with a finite rate of innovation (FRI). Classical sampling theory relies on signals being band-limited to enable perfect reconstruction; however, FRI signals—such as streams of Diracs or piecewise polynomials—are nonband-limited but still possess a finite number of degrees of freedom per unit of time. Shukla’s work, evidenced by a series of first-author publications, systematically expanded the mathematical tools available for sampling these signals in multiple dimensions.
In 2005, Shukla presented Sampling schemes for 2-D signals with finite rate of innovation using kernels that reproduce polynomials. This work addressed the challenge of sampling specific classes of 2-D FRI signals, namely sets of 2-D Diracs and bilevel polygons. Rather than relying exclusively on traditional sine or Gaussian kernels, Shukla proposed utilizing any sampling kernel capable of reproducing polynomials. By combining the polynomial approximation properties of these kernels with techniques such as complex-moments and the annihilating filter method, the research established new pathways for accurate signal reconstruction.
Building directly on this foundation, Shukla expanded the scope of these sampling schemes in 2006 with Tomographic Approach for Sampling Multidimensional Signals with Finite Rate of Innovation. This paper introduced a Radon transform-based hybrid scheme designed to handle more complex 2-D FRI signals, including piecewise polynomials with polygonal boundaries, as well as higher-dimensional Diracs and bilevel polytopes. A central technical contribution of this work was the introduction of an annihilating-filter-based-back-projection (AFBP) algorithm, which provided a robust mechanism for multidimensional signal recovery.
This sequence of conference papers culminated in a comprehensive journal publication in 2007, Sampling Schemes for Multidimensional Signals With Finite Rate of Innovation, published in the IEEE Transactions on Signal Processing. As the first author, Shukla consolidated the theoretical framework for sampling nonband-limited multidimensional signals. The paper formalized the argument that, despite falling outside the bounds of classical band-limited sine sampling theory, these signals can be perfectly reconstructed using a rich and flexible class of sampling kernels.
Current Directions
The most recent paper in the provided evidence is from 2007, representing the culmination of Shukla’s observed research on finite rate of innovation sampling. Based on this final documented work, the primary research direction involves the formalization of generalized sampling schemes for multidimensional nonband-limited signals. The core problem addressed in this most recent phase is the mathematical proof and practical algorithmic implementation required to perfectly reconstruct signals—such as streams of Diracs and piecewise polynomials—by exploiting their finite degrees of freedom. The work emphasizes the shift away from classical band-limited constraints toward the use of versatile polynomial-reproducing kernels in advanced signal processing applications.
Profile updated
Loading new papers…
Loading research neighbors…
See something wrong? Report a mistake