Muhammad Asad Lodhi

Papers

Total Citations

H-Index

About

Muhammad Asad Lodhi is a researcher whose work sits at the intersection of 3D computer vision and deep learning, with a primary focus on scene flow estimation from point clouds. His most significant contribution is the development of **FESTA (Flow Estimation via Spatial-Temporal Attention)**, a novel framework that addresses the critical challenge of understanding 3D scene dynamics—the movement and deformation of objects over time—directly from irregular, sparse point cloud data. This is a fundamental problem for applications like autonomous driving, robot navigation, and augmented reality, where traditional methods rely on dense RGB video. Lodhi’s approach leverages spatial-temporal attention mechanisms to capture long-range dependencies in the data, enabling more accurate and robust flow estimation than prior methods. His work on FESTA has garnered over 40 citations, underscoring its impact on the field and its adoption as a reference point for subsequent research. By pioneering attention-based architectures for 3D motion understanding, Lodhi is helping to bridge the gap between raw sensor data and high-level scene comprehension, paving the way for safer and more intelligent autonomous systems.