Research

My PhD research focus has been on rendering/inverse rendering and 3D reconstruction by exploiting the capabilities of deep learning-based AI techniques, as well as generative AI. My research has application in virtual reality, digital entertainment and multimedia.

In particular, I proposed a novel method of solving the rendering equation using a single neural network in a paper called Neural Radiosity. My follow-up papers(1, 2, 3) extended the idea of Neural Radiosity to 3D reconstruction, to account for global illumination in inverse rendering. By accounting for global illumination, my research enabled 3D reconstructions of the real world, in the wild (4). These projects were in line with my thesis title 'Towards inverse rendering with global illumination'. On a different topic, I proposed a method on importance sampling of neural incident radiance fields to accelerate rendering using path guiding. The resulting method is a US patent and is used in production in Disney company.

Another focus of my research has been on exploiting the capabilties generative AI for reconstruction tasks. In particular, I have upcoming papers on generative material details using diffusion models, in which I made innovations on training-free and multi-view consistent image generation. Some in progress projects involve augmenting NeRF-like training data using diffusion models as well as relighting textures.

Teaching

Since 2019, I have been TAing multiple intro/advanced computer graphics courses as well as other topics such as computer networks and data science.

Student Organizations

I have been an active member in advocacy student organizations on campus. In Fall 2021, I got elected as a CS representtative of Graduate Student Government (GSG). I also served as a member of International Student Advisory Board (ISAB) to consult our International Office (ISSS) throughout 2021-2023. In 2020, I served as a board member of Iranian Graduate Student Assiosiation (IGSF).