LAT: Local area transform for cross modal correspondence matching

详细信息    查看全文

• 作者：Seungchul Ryu ^{ryus01@yonsei.ac.kr}Author Vitae ; Seungryong Kim ^{srkim89@yonsei.ac.kr}Author Vitae ; Kwanghoon Sohn ; ^{khsohn@yonsei.ac.kr}Author Vitae
• 关键词：Cross-modal correspondence matching ; Local area transform ; Modality invariant ; Stereo matching
• 刊名：Pattern Recognition
• 出版年：2017
• 出版时间：March 2017
• 年：2017
• 卷：63
• 期：Complete
• 页码：218-228
• 全文大小：3250 K
• 卷排序：63

文摘

This paper describes a simple but powerful image tranform called local area transform (LAT) for modality-robust correspondence estimation. LAT transforms an image from the intensity domain to the local area domain, which is invariant under nonlinear intensity deformation, especially radiometric, photometric, and spectral deformations. Experimental results show that LA-transformed images provide consistency for nonlinearly deformed input images, even under random intensity deformations.