### Abstract

We present an adaptive neural network processor for image compression based on a modified frequency-sensitive self-organization algorithm. In this algorithm updating the codevector has a complexity of O(1) and O(N) for best case and worst case situations respectively. Experiments have shown that the worst case situation occurs only at the beginning stage of the learning process. The performance improves as the learning continues. Also the utilization of learning neurons has been considerably increased compared to other algorithm. This algorithm not only achieves a near-optimal result which is comparable with Linde-Buzo-Gray (LBG), but also retains the simplicity for hardware implementation. A mixed-signal architecture is proposed for this algorithm. It consists of analog circuitry which is responsible for neutral network computation and digital circuitry for frequency updating and losers selection.

Original language | English |
---|---|

Pages | 418-424 |

Number of pages | 7 |

Publication status | Published - 1994 Dec 1 |

Event | Proceedings of the 1994 IEEE International Workshop VLSI Signal Processing - La Jolla, CA, USA Duration: 1994 Oct 26 → 1994 Oct 28 |

### Other

Other | Proceedings of the 1994 IEEE International Workshop VLSI Signal Processing |
---|---|

City | La Jolla, CA, USA |

Period | 94-10-26 → 94-10-28 |

### Fingerprint

### All Science Journal Classification (ASJC) codes

- Signal Processing

### Cite this

*VLSI architecture for modified frequency sensitive self-organizing neural network for image data compression*. 418-424. Paper presented at Proceedings of the 1994 IEEE International Workshop VLSI Signal Processing, La Jolla, CA, USA, .