Title :
Temporal phase congruency
Author :
Myerscough, Peter J. ; Nixon, Mark S.
Author_Institution :
Dept. of Electron. & Comput. Sci., Southampton Univ., UK
Abstract :
We describe a robust moving feature detector that extracts feature points and feature velocities from a sequence of images. We develop a new approach based on phase congruency to include interpolation of feature orientation and improvements in robustness due to correlations in the image sequence. This new temporal phase congruency operator shows improved capabilities on a series of different real image types, as well as a noise analysis on synthetic images.
Keywords :
Gaussian noise; feature extraction; image sequences; interpolation; Gaussian noise; feature orientation interpolation; feature velocities; image sequence; moving feature detector; moving feature extraction; salt and pepper noise; temporal phase congruency; Bandwidth; Computer vision; Detectors; Digital filters; Frequency; Image edge detection; Image sequences; Lighting; Phase detection; Tuning;
Conference_Titel :
Image Analysis and Interpretation, 2004. 6th IEEE Southwest Symposium on
Print_ISBN :
0-7803-8387-7
DOI :
10.1109/IAI.2004.1300948
