Quantizing feature vectors in decision-making applications
US8271278B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Apr 3, 2010 |
| Grant date | Sep 18, 2012 |
| Priority date | — |
| Expiry date | Oct 15, 2030 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG10L17/06
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A system, method and computer program product for classification of an analog electrical signal using statistical models of training data. A technique is described to quantize the analog electrical signal in a manner which maximizes the compression of the signal while simultaneously minimizing the diminution in the ability to classify the compressed signal. These goals are achieved by utilizing a quantizer designed to minimize the loss in a power of the log-likelihood ratio. A further technique is described to enhance the quantization process by optimally allocating a number of bits for each dimension of the quantized feature vector subject to a maximum number of bits available across all dimensions.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.