3D pointcloud splatting (hot)
Novel splatting & rendering for dense urban point clouds — speed + fidelity tradeoffs tuned for city-scale reconstruction.
Gurudev · PhD (DSE), IISER Bhopal
GeoAI4Cities · 3D Reconstruction • LiDAR • Photogrammetry • Urban Digital Twins
Research & Focus
PhD candidate — Data Science & Engineering · GeoAI4Cities (IISER Bhopal)
I build GeoAI methods for urban-scale 3D perception: fusing LiDAR, UAV/satellite imaging, and multi-view photogrammetry to produce robust 3D reconstructions and semantic digital twins. My work combines geometry, deep learning, and scalable pipelines that target real city data and decision-making for urban resilience.
LiDAR + imagery fusion
End-to-end pipelines for fusing airborne LiDAR and multi-view imagery into textured building models and street-level twins.
Digital twin for resilience
Procedures to add attributes (materials, heights, emissions proxies) to 3D models for policy-use cases and simulation.
Urban semantic segmentation
Point-cloud segmentation with attention-aware networks for building, tree, and infrastructure classification.
What I work on
LiDAR Processing
Point-cloud denoise, registration, outlier removal, and attribute aggregation for city-scale scans.
Photogrammetry & Multi-view
Photogrammetric reconstruction, structure-from-motion, and hybrid fusion with depth sensors.
Deep Learning for 3D
Graph/Point networks, volumetric splatting, and multimodal feature learning for semantic understanding.