Cross Ratio-Based Refinement of Local Features for Building Recognition

Authors:
Hoang-Hon Trinh;Dae-Nyeon Kim;Kang-Hyun Jo
Affiliations:
Graduate School of Electrical Engineering, University of Ulsan, Korea, Ulsan, Korea 680 - 749;Graduate School of Electrical Engineering, University of Ulsan, Korea, Ulsan, Korea 680 - 749;Graduate School of Electrical Engineering, University of Ulsan, Korea, Ulsan, Korea 680 - 749
Venue:
ICIC '08 Proceedings of the 4th international conference on Intelligent Computing: Advanced Intelligent Computing Theories and Applications - with Aspects of Theoretical and Methodological Issues
Year:
2008

Citing 4
Cited 0

Using Discriminant Eigenfeatures for Image Retrieval

IEEE Transactions on Pattern Analysis and Machine Intelligence
Multiple View Geometry in Computer Vision

Multiple View Geometry in Computer Vision
Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
Hierarchical building recognition

Image and Vision Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper describes an approach to recognize buildings. The characters of building such as facets, their area, vanishing points, wall histogram and a list of local features are extracted and then stored in a database. Given a new image, the facet with biggest area is compared against the database to choose the closest pose. Novel methods of cross ratio-based refinement and SVD (singular value decomposition) based method are used to increase the recognition rate, increase the number of correspondences between image pairs and decrease the size of database. The proposed approach has been performed with 50 interest buildings containing 1050 images and a set of 50 test images. All images are taken under general conditions like different weather, seasons, scale, viewpoints and multiple buildings. We obtained 100(%) recognition rate.