GCH 2024 - Eurographics Workshop on Graphics and Cultural Heritage

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/3607042

Browse

Browsing GCH 2024 - Eurographics Workshop on Graphics and Cultural Heritage by Subject "Artificial intelligence"

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    DeepTex: Deep Learning-Based Texturing of Image-Based 3D Reconstructions
    (The Eurographics Association, 2024) Neumann, Kai Alexander; Santos, Pedro; Fellner, Dieter W.; Corsini, Massimiliano; Ferdani, Daniele; Kuijper, Arjan; Kutlu, Hasan
    Image-based 3D reconstruction is a commonly used technique for measuring the geometry and color of objects or scenes based on images. While the geometry reconstruction of state-of-the-art approaches is mostly robust against varying lighting conditions and outliers, these pose a significant challenge for calculating an accurate texture map. This work proposes a deep-learning based texturing approach called ''DeepTex'' that uses a custom learned blending method on top of a traditional mosaic-based texturing approach. The model was trained using a custom synthetic data generation workflow and showed a significantly increased accuracy when generating textures in the presence of outliers and non-uniform lighting.
  • No Thumbnail Available
    Item
    SemanticSplatStylization: Semantic scene stylization based on 3D Gaussian splatting and class-based style transfer
    (The Eurographics Association, 2024) Sinha, Saptarshi Neil; Graf, Holger; Weinmann, Michael; Corsini, Massimiliano; Ferdani, Daniele; Kuijper, Arjan; Kutlu, Hasan
    We propose a novel approach for 3D Semantic Style Transfer in 3D Gaussian Splatting (3DGS) that applies style transfer to specific segments of a 3D scene using 2D style images. Our method leverages a finetuning of 3D Gaussian splats and fast 2D class-based style transfer to achieve targeted stylization with superior fidelity and multi-view consistency compared to existing state-of-the- art methods. By incorporating a semantic understanding, our approach ensures precise, context-aware stylization, aligning the visual characteristics of segments with their intended style. The application of 3D Semantic Style Transfer in cultural heritage preservation and restoration holds significant potential. By accurately capturing and transferring styles onto specific segments of cultural heritage objects, our approach demonstrates the potential of providing more accurate and visually appealing stylization results that preserve the integrity and historical significance of cultural heritage artifacts.