Kyle Sargent

Papers

Total Citations

H-Index

About

Kyle Sargent is a researcher at the forefront of computer vision and graphics, specializing in novel view synthesis and generative modeling for dynamic scenes. His most cited work, "Generative Camera Dolly: Extreme Monocular Dynamic Novel View Synthesis" (2024), introduces a groundbreaking approach that enables photorealistic, temporally consistent video generation from a single monocular input. By leveraging generative priors, this method achieves extreme viewpoint changes—effectively creating a "dolly shot" effect—without requiring multi-view training data or explicit 3D reconstruction. This contribution addresses a long-standing challenge in the field, bridging the gap between static scene rendering and dynamic, free-viewpoint video. With 9 citations already in its first year, the paper signals strong early impact and relevance. Sargent’s work pushes the boundaries of what is possible with limited input, offering practical tools for filmmakers, VR content creators, and robotics. His research exemplifies how generative AI can transform traditional graphics pipelines, making high-quality dynamic view synthesis accessible and efficient. For students and researchers, Sargent’s approach represents a compelling fusion of learning-based and geometric methods, opening new avenues for immersive media and autonomous perception.