A self-organized clustering system is extensively used for neural networks for the analysis and classification of speech and image information data. The speed of this process is improved by applying a new parallel algorithm essential for a real system and the recognition rates achieved by this new system is almost same as that of a sequential system using the conventional clustering system. However, the use of a multiple number of processors reduces the total processing time by a fraction.
|Number of pages||16|
|Journal||Neural Network World|
|Publication status||Published - 1996|
All Science Journal Classification (ASJC) codes
- Hardware and Architecture
- Artificial Intelligence