In this work, we presented the worked out algorithms of digital filtration in mixed and combined formats. We use these formats to improve the speed and economical of worked out algorithms. To find an application for the above-mentioned algorithm in neural networks, increases and DM-steps of input signal and impulse response were presented in low bits binary codes. On the basis of these approaches, the corresponding convolutions were realized by use only the simple logical operations on the codes of input signal and impulse response.
Convolutions in mixed formats oriented on multiprocessors systems
2003-01-01
205420 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Convolutions in Mixed Formats Oriented on Multiprocessors Systems
British Library Conference Proceedings | 2003
|Highly reliable multiprocessors
NTRS | 1977
|Data Formats: Using Self-Describing Data Formats
NTIS | 2011
|Image Inpainting with Onion Convolutions
British Library Conference Proceedings | 2021
|Shared versus distributed memory multiprocessors
TIBKAT | 1991
|