Projects
Easy and flexible deployment of custom deep convolutional neural networks.Â
Documentation: https://dlsia.readthedocs.io/en/latest/
Arxiv preprint: https://arxiv.org/pdf/2308.02559.pdf
Image inpainting is a restoration process that estimates the contents of missing regions within images and videos. Here, we train deep learning models to recover information missing behind detector gaps.
I deploy deep learning models for learning a low dimensional representation of data.
Link: https://dlsia.readthedocs.io/en/latest/tutorialLinks/latentSpaceExploration_SMSNetAutoencoders.html
Data curation and labeling is time consuming... to that end, I leverage multi-network aggregation and random network topologies to alleviate the data-intensive requirements of supervised deep learning segmentation.
My PhD was spent investigating topology and fluid dynamics through the lens of computational geometry.