Optics, Quantum & CV Lab
At Tel Aviv University’s Optics, Quantum & Computer Vision Lab, I worked on hyperspectral imaging systems and machine learning pipelines for image analysis manipulation.
My research included training and evaluating models for spectral and image-based tasks, as well as building experimental imaging setups using Fresnel and liquid lenses. A primary focus of my work was exploring practical and experimental use cases of Fresnel optics in compact imaging systems.
In parallel, I collaborated with an early-stage startup on data collection for facial recognition systems, supporting controlled data acquisition through participant recruitment and experimental trials to generate real-world datasets for model development and validation.
Research Contributions & Projects
Joined ongoing hyperspectral imaging project and contributed to system training, data collection, and optical refinement. As part of the system redesign, helped implement a Fresnel Lens based optical configuration that reduced sensor lens spacing by 7mm, enabling a more compact design
Hyper Spectral Imaging System
Applied CycleGAN-based unpaired image translation to correct distortion in Fresnel-lens imaging systems. This approach leverages machine learning to compensate for optical artifacts without paired data, enabling lighter and more compact camera designs.
Unpaired Image to Image CycleGAN
Contributed to data collection, curation, and training of machine learning models to improve image reconstruction quality across multiple imaging pipelines and projects. This work enabled comparative evaluation of optical configurations and improved reconstruction robustness for both experimental and applied imaging systems.
Machine learning training
Led end-to-end data collection for a facial recognition dataset involving 120+ participants over a three-week trial. Designed and executed the collection protocol, coordinated participant recruitment and consent, managed daily system setup and teardown, and oversaw secure data upload and validation to ensure dataset quality and consistency.