Rishi Upadhyay

I am a 2nd year PhD student at UCLA where I am advised by Prof. Achuta Kadambi. My research interests include computer vision and computer graphics, particularly diffusion models.

I attended UC Berkeley for my undergrad where I was fortunate to work with Prof. Ren Ng and Prof. Avideh Zakhor.

Email / Google Scholar / Github

Publications

Teaching

Representative/Recent papers are highlighted
	WeatherProof: A Paired-Dataset Approach to Semantic Segmentation in Adverse Weather Blake Gella, Howard Zhang, Rishi Upadhyay, Tiffany Chang, Matt Waliman, Yunhao Ba, Alex Wong, Achuta Kadambi arXiv, 2023 PDF / project page We introduce the first paired semantic segmentation dataset in adverse conditions and show that techniques that leverage this paired nature can significantly improve performance.
	SparseGS: Real-Time 360° Sparse View Synthesis using Gaussian Splatting Haolin Xiong, Sairisheek Muttukuru, Rishi Upadhyay, Pradyumna Chari, Achuta Kadambi arXiv, 2023 PDF / project page We integrate 3D Gaussian Splatting with depth regularization and diffusion priors to enable 360° novel view synthesis from just 12 views.
	Enhancing Diffusion Models with 3D Perspective Geometry Constraints Rishi Upadhyay, Howard Zhang, Yunhao Ba, Ethan Yang, Blake Gella, Sicheng Jiang, Alex Wong, Achuta Kadambi ACM TOG, SIGGRAPH Asia, 2023 PDF / Code / Project Page We propose a loss function for latent diffusion models that improves the perspective accuracy of generated images, allowing us to create synthetic data that helps improve SOTA monocular depth estimation models.
	Characterizing Cone Spectral Classification by Optoretinography Vimal Prabhu Pandiyan , Sierra Schleufer, Emily Slezak, James Fong, Rishi Upadhyay, Austin Roorda, Ren Ng, Ramkumar Sabesan Biomed. Opt. Express, 2022 PDF We characterize cone classification by ORG. Personally worked on the Retina Map Alignment Tool.
	Scaling Building Energy Audits through Machine Learning Methods on Novel Drone Image Data Reshma Singh, Samuel Fernandes, Anand Krishnan Prakash, Paul Mathew, Jessica Granderson, Colman Snaith, Rahul Pusapati, Prakash Jadhav, Avideh Zakhor, Rishi Upadhyay, Ozgur Gonen, Harry Bergmann Technical Report, Lawrence Berkeley National Labratory, 2022 PDF We create an automated tool to estimate building energy efficiency through drone and thermal imaging.
	Indoor 3D Interactive Asset Detection Using a Smartphone Revekka Kostoeva, Rishi Upadhyay, Yersultan Sapar, Avideh Zakhor ISPRS, 2019 PDF We use a smartphone, AR techniques, and object classifiers to create a system to automatically detect and map objects of interest in a 3D scene.

Credit to Jon Barron for the website template.