Fast calculation of correlations in recognition systems

TL;DR

This paper introduces a computationally efficient classification architecture that leverages a fast tensor-vector multiplication algorithm, applicable across various recognition systems from simple filters to complex neural networks.

Contribution

It presents a novel fast tensor-vector multiplication method that enhances the efficiency of recognition system computations across diverse architectures.

Findings

Significant reduction in computation time for recognition tasks.

Applicable to both simple and complex recognition systems.

Potential for improved real-time recognition performance.

Abstract

Computationally efficient classification system architecture is proposed. It utilizes fast tensor-vector multiplication algorithm to apply linear operators upon input signals . The approach is applicable to wide variety of recognition system architectures ranging from single stage matched filter bank classifiers to complex neural networks with unlimited number of hidden layers.