Three dimensional models from microscopy images using Gaussian splatting

Authors

Khristian Kikuchi ⋅ PH Data Science Program, College of Science, University of the Philippines Diliman and College of Computer and Information Science, Mapúa Malayan Colleges, Laguna
Patricia Cabrera ⋅ PH School of Archaeology, University of the Philippines Diliman
Jayson Victoriano ⋅ PH College of Information and Communication Technology, Bulacan State University
Juan Rofes ⋅ PH School of Archaeology, University of the Philippines Diliman and Archéozoologie, Archéobotanique Sociétés, Pratiques et Environnements, CNRS/MNHN, France and National Museum of the Philippines
Giovanni Tapang ⋅ PH National Institute of Physics, University of the Philippines Diliman

Abstract

We introduce a novel approach to constructing 3D models of small objects, specifically bone elements and fragments, using Gaussian splatting. We compare a reconstructed 3D model to the original object using OpenSplat to implement Gaussian Splatting and generate the 3D model for a 10-peso Philippine coin. We show a comparative analysis of selected recognizable features of the coin. Results demonstrate the feasibility of our approach, as we were able to construct a 3D model of an object using 32 input views with recognizable features. We present the reconstructed 3D model of a rodent bone fragment as an application.

Downloads

Issue

2024: Proceedings of the 42nd Samahang Pisika ng Pilipinas Physics Conference

Brewing waves of innovation and discovery in Physics
3-6 July 2024, Batangas State University, Pablo Borbon Campus

Please visit the SPP2024 activity webpage for more information on this year's Physics Congress.

Article ID

SPP-2024-3C-05

Section

Optics and Photonics

Published

2024-06-29

How to Cite

[1]

K Kikuchi, P Cabrera, J Victoriano, J Rofes, and G Tapang, Three dimensional models from microscopy images using Gaussian splatting, Proceedings of the Samahang Pisika ng Pilipinas 42, SPP-2024-3C-05 (2024). URL: https://proceedings.spp-online.org/article/view/SPP-2024-3C-05.