Projects

The Panning and Translational Invariant Incremental Principal Component Pursuit (incPCP-PTI) algorithm is a newly developed fully incremental method for video background modeling that extends incPCP-TI in order to cope with moving and panning cameras. The method continuously updates the low-rank component in order to align it to the current reference frame of the camera. To the best of our knowledge, this algorithm is the first low rank plus additive matrix algorithm capable of handling both panning and jitter. This project was performed with the Digital Signal Processing Laboratory at PUCP.

The minimum variance distortionless response (MVDR) beamformer is an adaptive beamformer that can achieve superior resolution compared to conventional delay-and-sum (DAS) beamforming. The objective of the project was to first study the effects of aberration and conventional methods of aberration correction on the MVDR Beamformer. In the second phase of the project, we arrived to a novel locally adaptive phase aberration correction method that performs better than previous methods for both DAS and MVDR. This project is a collaboration between the Medical Imaging Laboratory at PUCP and the Dahl Ultrasound Research Group at Stanford University . 

Subvocal speech recognition aims at recognizing EMG signals from the neck area generated when a person utters words without necessarily producing vocal speech. Our objective was to correctly classify a limited set of phrases using only one EEG channel using cumulative residual entropy features and a support vector machine classifier. This project corresponded to my undergraduate thesis project.