Article,
Efficient Fuzzy Set Theoretic Approach to Image Corner Matching
International Journal on Recent and Innovation Trends in Computing and Communication, 3 (2): 508--511 (February 2015)
DOI: 10.17762/ijritcc2321-8169.150217
Abstract
Corner matching in digital images is a key step in several applications in computer vision such as motion estimation, object recognition and localization, 3D reconstruction etc. Accuracy and reliability of corner matching algorithms are two important criteria. In this paper, corner correspondence between two images is established in the presence of intensity variations, inherent noise and motion blur using a fuzzy set theoretic similarity measure. The proposed matching algorithm needs to extract set of corner points as candidates from both the frames. Fuzzy set theoretic similarity measure is used here as fuzzy logic is a powerful tool which is able to deal with ambiguous data. Experimental results conducted with the help of various test images show that the proposed approach is superior compared existing corner matching algorithms (conventional and recent) considered in this paper under non-ideal conditions
