I come from signal processing background with emphasis on image processing. I've also taken courses in machine
learning and control. I want to combine to find new and exciting applications of computer vision
and image processing in robot perception.
Explored rendering equation optimization and machine learning based methods to estimate scene parameters
such as depth and surface reflectance based on a single image.
Designed, built, and tested the electronics that integrates the pod's computer, sensors, actuators,
communication system, and battery system that controlled the first linear induction motor propelled
hyperloop in the SpaceX competition. Also designed, built and tested the wiring harness that can withstand the
vibrational load.
Implemented an Octree-accelerated path tracer with multiple importance sampling, various BSDF (including Disney BSDF), various
lighting (including image based lighting) for CS-440 Advanced Computer Graphics at EPFL.
Digital Lippmann Camera is a hyperspectral camera concept developed at LCAV. In order to recover the spectral
information of the scene, the camera requires sub-micron precision control of the mirrors in a Michelson interferometer.
This project utilized monochromatic lasers and photodiodes to find accurate displacement of piezo actuators
for controlling the mirrors of the imager.
This project explores various ways of detectingi the beat of the music, and automatically control a Neopixel
LED matrix to time the light changes with the beat of the music. The various methods (Energy, STFT, etc.) are
tested on several audio clips, and their performances are compared.
Designed new Jupyter Notebook exercises to introduce students to the concept of Fourier Series, Fourier Transform, sampling,
as well as important Python practices in numerical coding such as vectorization and masking. Also led two discussion/lab section
every other week for a total of 6 sessions during the semester, both in person and remote.