Samrat Dutta
Papers
1
Total Citations
7
H-Index
1
About
Samrat Dutta is a researcher whose work lies at the intersection of computer vision and deep learning, with a particular focus on monocular depth estimation. His most-cited paper, "Unsupervised Depth and Confidence Prediction from Monocular Images using Bayesian Inference" (2020), introduces an innovative unsupervised deep learning framework that leverages Bayesian inference to enhance per-pixel depth prediction from single RGB images. A key contribution of this work is the simultaneous prediction of confidence maps alongside depth and pose information, enabling more reliable and interpretable 3D scene understanding. This approach addresses a critical challenge in autonomous systems—how to assess the certainty of depth predictions without ground truth data. With 7 citations, Dutta’s research has already begun influencing the field, offering a robust foundation for future work in self-supervised learning and uncertainty quantification. His achievements demonstrate a commitment to advancing practical, data-efficient solutions for real-world applications like robotics and augmented reality, where accurate depth perception is essential.
Research Focus
Key Achievements
Top Papers
- 1