Analysis of efficient linear local features of digital signals and images

Abstract

The paper presents the analysis of the efficiency of two original approaches to the construction of sets of linear local features (LLF) of digital signals. The first approach is based on the construction of the LLF set of separately constructed efficient LLFs, each of which has its own algorithm for the feature computation. The second approach involves the construction of an efficient LLF set that has a single algorithm for the simultaneous computation of all features. The analysis is carried out with respect to several indicators that characterize computing and qualitative properties of constructed LLFs. The two considered approaches are also experimentally compared with known solutions.