Saved in:
Bibliographic Details
Main Authors: Nam, Myeongseok, Park, Wongi, Kim, Minsol, Hur, Hyejin, Lee, Soomok
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.19138
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910967522131968
author Nam, Myeongseok
Park, Wongi
Kim, Minsol
Hur, Hyejin
Lee, Soomok
author_facet Nam, Myeongseok
Park, Wongi
Kim, Minsol
Hur, Hyejin
Lee, Soomok
contents Recently, 3D Gaussian Splatting (3D-GS) based on Thermal Infrared (TIR) imaging has gained attention in novel-view synthesis, showing real-time rendering. However, novel-view synthesis with thermal infrared images suffers from transmission effects, emissivity, and low resolution, leading to floaters and blur effects in rendered images. To address these problems, we introduce Veta-GS, which leverages a view-dependent deformation field and a Thermal Feature Extractor (TFE) to precisely capture subtle thermal variations and maintain robustness. Specifically, we design view-dependent deformation field that leverages camera position and viewing direction, which capture thermal variations. Furthermore, we introduce the Thermal Feature Extractor (TFE) and MonoSSIM loss, which consider appearance, edge, and frequency to maintain robustness. Extensive experiments on the TI-NSD benchmark show that our method achieves better performance over existing methods.
format Preprint
id arxiv_https___arxiv_org_abs_2505_19138
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Veta-GS: View-dependent deformable 3D Gaussian Splatting for thermal infrared Novel-view Synthesis
Nam, Myeongseok
Park, Wongi
Kim, Minsol
Hur, Hyejin
Lee, Soomok
Computer Vision and Pattern Recognition
Recently, 3D Gaussian Splatting (3D-GS) based on Thermal Infrared (TIR) imaging has gained attention in novel-view synthesis, showing real-time rendering. However, novel-view synthesis with thermal infrared images suffers from transmission effects, emissivity, and low resolution, leading to floaters and blur effects in rendered images. To address these problems, we introduce Veta-GS, which leverages a view-dependent deformation field and a Thermal Feature Extractor (TFE) to precisely capture subtle thermal variations and maintain robustness. Specifically, we design view-dependent deformation field that leverages camera position and viewing direction, which capture thermal variations. Furthermore, we introduce the Thermal Feature Extractor (TFE) and MonoSSIM loss, which consider appearance, edge, and frequency to maintain robustness. Extensive experiments on the TI-NSD benchmark show that our method achieves better performance over existing methods.
title Veta-GS: View-dependent deformable 3D Gaussian Splatting for thermal infrared Novel-view Synthesis
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.19138