Fast calculation of correlations in recognition systems
This work addresses computational bottlenecks in recognition systems for applications requiring efficient classification, but it appears incremental as it builds on existing tensor methods without claiming major breakthroughs.
The authors tackled the problem of computational efficiency in recognition systems by proposing a new architecture that uses fast tensor-vector multiplication to apply linear operators on input signals, resulting in applicability to a wide range of systems from simple classifiers to complex neural networks.
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.