Vaibhav Kumar

Vaibhav Kumar Upadhyay, PhD

Assistant Professor

Data Science and Engineering

Head In Charge (Computer Center)

Sports Faculty Advisor

Indian Institute of Science Education and Research, Bhopal

B.E. : Computer Science and Engineering

M.Tech : Geoinformatics

PhD : Centre for Urban Science and Engineering, IIT Bombay

Research Interests:

Geospatial Artificial Intelligence (GeoAI), 2D/3D GIS, Urban Informatics.

ORCID, Google Scholar, ResearchGate

PI: GeoAI4Cities Lab

Advisor: SimDAAS Pvt. Ltd.

Teaching

Artificial Intelligence (DSE313/ECS 313)

Spatial Data Science and Applications (DSE 406/606)

3D Deep Learning and Applications (DSE411/611)

Robotic Perception (DSE425/DSE625)

If you are passionate about 3D Vision, Urban Planning, and various verticals of automating Digital Twins (Development and Application) using AI, please Leave A Message or email me if you are interested.

Hot from the Lab!!: First of its kind, Freely available labeled LiDAR data repository from Indian Cities. Access it at :https://www.lidaverse.com/#/

RECENT PUBLICATIONS

Chauhan, P.L., Bais, A.S. & Kumar, V. (2026). Performance analysis of subsampled LiDAR point clouds using deep learning based semantic segmentation. Appl Intell 56, 273 . https://doi.org/10.1007/s10489-026-07282-2
Pratap, B., & Kumar, V. (2026). CVPC: Cross-Modal Visual-Guided Point Cloud Completion. IEEE Robotics and Automation Letters. https://doi.org/10.1109/LRA.2026.3683588
Kumar, V., et al. (2026). Ke-MLS: A knowledge-enhanced mobile laser scanning dataset for urban scene understanding. Environment and Planning B: Urban Analytics. https://doi.org/10.1177/23998083261430812

My research focuses on developing smart digital twins for urban environments, integrating geospatial AI, remote sensing, and 3D data science. The work is structured around three major verticals.

I focus on twin attribution and feature extraction using high-resolution remote sensing data, including imagery and LiDAR. This involves semantic segmentation, object detection, and attribution of key urban features critical for downstream planning and analysis.

I work on the development of digital twins using learning and optimization techniques, particularly deep learning applied to 3D representations such as point clouds, meshes, and depth maps. My research explores accurate 3D data generation, as well as transformations across different 3D formats to enable interoperability and flexibility in simulation environments.

I emphasize synthetic 3D representation generation using multimodal data sources. This includes the automatization of large-scale outdoor 3D maps using deep generative models trained on aerial, street-view, and ground-truth data.

I apply these smart digital twins to diverse use cases, including urban planning, climate resilience, and simulation environments for autonomous systems. My goal is to build scalable, transferable models that support intelligent decision-making in complex urban settings.