Skip to content
Gyanateet Dutta
Work·Academic·CV

Work

2024Systems

Dalton Mills VR Reconstruction

Collaborator, Science Museum Group / Leeds HELIX XR

From September to December 2024, I worked with the Science Museum Group’s Congruence Engine and University of Leeds HELIX XR on a VR reconstruction of Dalton Mills in Keighley, which was destroyed by fire in 2022. I collaborated with Alex Neish, Yuan Gao, and Simon Popple. The PCVR reconstruction used photogrammetry and NeRF methods on a small collection of photographs.

A point-cloud mill facade with ten camera positions and colour-coded photographed and unphotographed surfaces.
Fig. 1A reconstruction schematic showing the input viewpoints and the surfaces they cover. Camera positions and coverage are illustrative.
Scope

Geometric fidelity follows the surviving photograph set; no pre-fire laser scan exists for comparison.

Stack
Unreal Engine 5, NeRF, Photogrammetry, VR
Links
Leeds case study, Congruence Engine, Study repo

Also

  • 2026CUDA Blackwell Labs
  • 2026Agentic structure-from-motion
  • 2026PCOS edge agent
  • 2023AWS AI/ML Scholar
  • B3tt3r: 3D Reconstruction
GitHub· ORCID· Google Scholar· LinkedIn· Hugging Face· X