상세정보
학술대회
High Quality Texture Reconstruction from Multi-view Images
- 발행일
- 201710
- 출처
- International Conference on Information and Communication Technology Convergence (ICTC) 2017, pp.1113-1115
- 협약과제
- 17HS1400, 다고도 이기종 실사영상 기반 3차원 합성전장 구축 및 가시화 기술 개발, 박창준
- 초록
- Texture reconstruction is the last step of 3D reconstruction pipelines that makes 3D geometry models very impressive and beautiful. Once the 3D model is created through structure-from-motion, multi-view stereo and mesh reconstruction, real image based texture is mapped to the models. But it is not easy to generate high quality texture. There are many difficulties cause problems: the large number of input images, their exposure variations, and external illuminance change. The proposed technique is able to generate seamless and high quality textures by view selection optimization and color adjustment.
- 키워드
- 3D reconstruction, texture map generation, Texture reconstruction
- KSP 제안 키워드
- 3D Reconstruction, 3D geometry, 3d model, Color adjustment, Map generation, Mesh reconstruction, Multi-view stereo, Texture map, View Selection, multi-view images, structure from motion