UTexGen: High‐quality texture reconstruction for large‐scale scenes using multi‐view images

• DOI: 10.4218/etrij.2024-0320
• Publication Date: 2024-12-12T04:33:59Z
• Authors: Hye‐Sun Kim; Yun‐Ji Ban; Chang‐Joon Park
• Keywords: Texture filtering; Leverage (statistics); Texture (cosmology); Texture atlas

Abstract When reconstructing extensive terrain, it is essential to partition into smaller tiles for individual processing. This paper introduces a texture reconstruction approach that ensures seamless and consistent final outputs, even when processed tile by tile. Among the stages of multi‐view image‐based reconstruction, presents significant challenges during tile‐based Relying solely on local tile‐level data complicates achieving precise mapping. The absence occlusion details between can lead selecting incorrect images as best visible ones or adjusting colors differently, resulting in noticeable grid‐like seams result. To mitigate these issues, we leverage global depth maps accurately detect occlusions neighboring tiles. Furthermore, utilizing shared candidate list, establish uniform targets color correction across Experimental findings demonstrate leveraging information tile‐by‐tile basis enables creation smooth realistic maps, validated through comparisons with existing methodologies.