Karan Narain

Papers

Total Citations

H-Index

About

Karan Narain is a computer vision researcher whose work centers on efficient 3D pose estimation for resource-constrained environments. His most impactful contribution, "3DPoseLite: A Compact 3D Pose Estimation Using Node Embeddings" (2021, 15 citations), addresses a critical challenge in augmented reality (AR), autonomous navigation, and robotics: enabling accurate pose estimation directly on edge devices. By introducing a novel node embedding approach, Narain’s method achieves a compelling balance between model compactness and precision, making it suitable for real-time on-device inference where computational budgets are tight. This work has practical implications for deploying AR applications on mobile hardware and improving the autonomy of drones and robots. Beyond this flagship paper, Narain’s research continues to push the boundaries of efficient deep learning for 3D understanding, with a focus on bridging the gap between algorithmic accuracy and real-world deployment constraints. His contributions are particularly valuable for students and engineers seeking to build lightweight, high-performance vision systems for interactive and autonomous technologies.