2D image detection using state of the art models to detect trees
on raster data and optimize forest data collection
* Accepted as workshop paper @ ICLR 2025
Using multi-angled sparse views of an individual's face to create a 3D facial reconstruction with textured mapped tattoos.
A perceptually realistic cloth simulator with several shaders to
stylize a simulation of a piece of cloth colliding with various
objects.
Additionally, check out my other computer graphics projects
here.
Crowdsourced video diffusion evaluation platform allowing users to access, compare and rank text-to-video generative models.
Implemented and deployed diffusion models using various sampling loops to generate various images using different conditioning techniques (inpainting, optical illusions etc.). Additionally developed and trained a UNet using unconditional, time condition and class conditioned variants.
Developed an automatic correspondance detector used to warp and
morph corresponding facial features.
Additionally, check out my other computer vision projects
here.
Designed and implemented a client application for secure file sharing system using cryptographic primitives.
Customized automated instagram messaging by simulating user interaction with the Instagram API. Additionally sources sellers using hashtag scraping and backend processing to generate csv files of eligible usernames.
More to come...