A new version of the flusser invariants set for pattern feature extraction
Constantin-Iulian Vizitiu, Doru Munteanu · Computational intelligence · 2007
The choice of an appropriate feature extraction method is essential for the success of the following recognition/classification process. This paper proposes a design method for a new pattern descriptors set belonging to the Flusser moments family, which is invariant to the elementary geometric transformations and respectively, has an increased noise robustness. Experimental results based on a real video image database, confirm the basic properties of this new descriptors set.