Riu Cherdchusakulchai
Riu Cherdchusakulchai is a Master’s student in Computer Vision in the Robotics Institute at Carnegie Mellon University. He is advised by Prof. Fernando De la Torre, working in the Human Sensing Lab. Before that, He had three years working experience as a Machine Learning Engineer at AI and Robotics Ventures. He received my Bachelor’s degree in Mechanical Engineering from Chulalongkorn University, advised by Prof. Nattapol Damrongplasit. His research interests lie at the intersection of 3D vision, spatial intelligence, and robotics. He is particularly interested in neural scene reconstruction and 3D understanding for augmented reality and robotics applications.
Shreyans Sethi
Shreyans Sethi is a Master’s student in Computer Vision at Carnegie Mellon University. His background is in multimodal machine learning, including work at IBM Research on multimodal biomedical foundation models and current research on adapting vision-language models to out-of-domain tasks. He is especially interested in the intersection of computer vision, 3D data, and biomedical or healthcare applications. For this project, Shreyans primarily contributed to the motion-guided reconstruction pipeline, including computing optical flow from multi-view video, lifting 2D flow into 3D motion estimates, and exploring how optical flow can be used to regularize Gaussian updates.
Abhinav Mahajan
Abhinav Mahajan is a Master’s student in Computer Vision at Carnegie Mellon University. His background is in Computer Vision and multimodal machine learning, including work at Adobe Research on graphic design aesthetic evaluation and refinement and at Amazon on semantic personalized recommendation, with current research on 3D/4D vision for perception systems that can infer the spatial and physical structure of the world. For this project, Abhinav primarily contributed to the motion-guided reconstruction pipeline, the algorithm for lifting 2D flow into 3D motion estimates, and exploring the theoretical capabilities of the Hash Grids of the NTC and make it better and faster.
Liujie Zheng
Liujie Zheng is a Master’s student in Computer Vision at Carnegie Mellon University with experience in machine learning. His work spans from building production-grade LLM-powered systems, such as scalable voice chatbots and multi-agent platforms, to research in real-time 3D reconstruction and immersive telepresence. He enjoys bridging theory and practice by developing efficient, scalable AI systems for real-world impact. For this project, Liujie primarily worked on compression of 4D Gaussians with efficient data structures such as hash grids and octrees, and algorithms such as entropy coding.
Advisors
Our team is advised by Professor Fernando de la Torre, a research professor in the Robotics Institute at Carnegie Mellon University. He is also the director of the Human Sensing Lab, which this project falls under, and has many years of experience in the intersection of augmented reality, virtual reality, and computer vision/machine learning.
Saswat Subhajyoti Mallick, a research associate in the Robotics Institute at Carnegie Mellon, is also an advisor on this project. His interests lie in the intersection of 3D computer vision, spatial intelligence and graphics for robotics applications and immersive experiences. Having published highly relevant work in the domain of this project, Saswat is able to provide valuable guidance and insights on our ideas.