Sampling and Reconstruction of Warped Signals
Imagine taking a photo from a curved surface or having a record player that plays the distorts the records with some fixed but unknown function when playing them. Can we find the curvature of the photographed object or the original recoded sound in the above examples? This problem is known as reconstructing warped signals, and has many applications in simultaneous localization and mapping (SLAM), shape and depth estimation, MRI imaging, etc.
In this research we work on algorithms for reconstructing the warping function and original signal based on the given bandwidth of the original signal in the context of “shape from bandwidth”.
We also study the sampling and reconstruction of linearly warped periodic signals and propose an algorithm for efficiently estimate the warping and the original signal.
In this project we use cool mathematical tools to design 3D barcode systems and introduce algorithms to extract information from such barcodes.
Stay tuned to hear more about this soon …