الفهرس | Only 14 pages are availabe for public view |
Abstract Image moments and their invariants to translation, scaling and rotation transformations were successfully used in image processing and pattern recognition applications. In pattern recognition applications, moment invariants are used as invariant features that characterize the digital images. High accurate, fast and numerically stable computing is crucial for ensuring these moment invariants accuracy and stability where the extracted features used with classifiers to discriminate between similar images. The accuracy of the extracted features is the corner stone in the success of the classification process. It is well known that, accurate image reconstruction required higher order moments. Therefore, Accurate, fast and numerically stable computation of moments invariants are big challenges. Recently, several approaches were developed to deal with these challenges. |