On image analysis by orthogonal moments
- Already Purchased? View Article
- Subscription Options Learn More
Examines the reconstruction properties of orthogonal moment descriptors. An asymptotic expansion for the global reconstruction error is established. This reveals mutual relationships between a number of moments, the image smoothness, sampling rate and noise model characteristics. The problem of a data-driven selection of an optimal number of moments is addressed
Date of Conference: 30 Aug-3 Sep 1992