Aritra Mukherjee

Papers

Total Citations

H-Index

About

Aritra Mukherjee is a computer vision researcher whose work bridges classical statistical methods and modern deep learning, with a particular focus on scene understanding for autonomous systems. His research spans three key areas: simultaneous localization and mapping (SLAM), point cloud processing from LiDAR data, and semantic segmentation of natural images. Mukherjee’s most impactful contribution is his 2019 work, "Detection of loop closure in SLAM: A DeconvNet based approach," which has earned 24 citations. This paper introduced a novel deep learning framework for detecting loop closures—a critical challenge in SLAM that prevents drift in robotic navigation. He has also advanced geometric surface-based segmentation of LiDAR point clouds, enabling faster processing for autonomous vehicles. In semantic segmentation, Mukherjee has explored hybrid approaches, combining superpixel-based statistical methods with deep architectures like ForkNet, as seen in his two-stage segmentation work. His research consistently demonstrates a practical focus on improving computational efficiency while maintaining accuracy in diverse, natural environments. With publications spanning 2019-2020, Mukherjee’s work represents a thoughtful integration of traditional and contemporary techniques in visual perception.