Papers
2
Total Citations
10
H-Index
2
About
S P Sharan is a researcher making impactful contributions at the intersection of computer vision and deep learning, with a focus on efficient image and video processing. His work addresses the critical challenge of deploying high-performance deep learning models in real-world, resource-constrained environments. Sharan’s most cited paper, “Paced-curriculum distillation with prediction and label uncertainty for image segmentation” (2023, 7 citations), introduces a novel knowledge distillation framework that leverages curriculum learning and uncertainty estimation to improve segmentation accuracy, a fundamental task in medical imaging and autonomous driving. In “RepAr-Net: Re-Parameterized Encoders and Attentive Feature Arsenals for Fast Video Denoising” (2022, 3 citations), he tackles the practical need for real-time video denoising in mobile robotics, satellite television, and surveillance. This work demonstrates how re-parameterization and attention mechanisms can bridge the performance gap between traditional, lightweight methods and computationally heavy deep learning models, enabling superior denoising without sacrificing speed. By pioneering techniques that balance accuracy, efficiency, and real-time applicability, Sharan is advancing the frontier of deployable AI for visual perception tasks.
Research Focus
Key Achievements
Top Papers
- 1
- 2