Vector quantization using thresholds
First Claim
1. In a computer system, a method of vector quantization, comprising the following steps:
- a. from a training set of vectors, creating a vector quantization codebook representative of a set of training vectors, said codebook comprising a representative set of n quantized vectors from said training vectors;
b. for each i quantized vector (VQi) of said n quantized vectors, determining an i threshold (Thresholdi) to be associated with said i quantized vector, wherein said i threshold comprises half a distance between said i quantized vector and a k quantized vector (VQk) of said n quantized vectors, wherein i<
k≦
n, and said k quantized vector is the closest quantized vector to said i quantized vectors of a set of quantized vectors having an index k wherein i<
k≦
n;
c. for each i quantized vector of said n quantized vectors, associating in said vector quantization codebook said i threshold with said i quantized vector; and
d. using said vector quantization codebook to encode a set of input vectors (V1 -VTOT).
2 Assignments
0 Petitions
Accused Products
Abstract
Methods and apparatus for vector quantization. A threshold generator generates an i threshold (Thresholdi) to be associated with each i quantized vector of n quantized vectors in a vector quantization codebook. The vector quantization codebook and the thresholds are used by a vector quantizer to encode a set of input vectors (V1 -VTOT). The determination that a distance between a vector to be encoded and a quantized vector in a codebook is less than the associated threshold causes a search for the closest vector to terminate for a nearest neighbor vector quantizer. In some embodiments, the vectors comprise samples of continuous signals for sound containing speech, or display signals. In other embodiments, codebook vectors are arranged from most frequently encoded vectors to least frequently encoded vectors.
-
Citations
10 Claims
-
1. In a computer system, a method of vector quantization, comprising the following steps:
-
a. from a training set of vectors, creating a vector quantization codebook representative of a set of training vectors, said codebook comprising a representative set of n quantized vectors from said training vectors; b. for each i quantized vector (VQi) of said n quantized vectors, determining an i threshold (Thresholdi) to be associated with said i quantized vector, wherein said i threshold comprises half a distance between said i quantized vector and a k quantized vector (VQk) of said n quantized vectors, wherein i<
k≦
n, and said k quantized vector is the closest quantized vector to said i quantized vectors of a set of quantized vectors having an index k wherein i<
k≦
n;c. for each i quantized vector of said n quantized vectors, associating in said vector quantization codebook said i threshold with said i quantized vector; and d. using said vector quantization codebook to encode a set of input vectors (V1 -VTOT). - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9)
-
-
10. An apparatus for vector quantization comprising:
-
a. a vector quantization codebook generator for generating a representative set of n quantized vectors operative upon receiving a set of training vectors; b. a threshold generator operative for generating a threshold (Thresholdi) for each i quantized vector (VQi) of said n quantized vectors, said i threshold comprising half a distance between said i quantized vector and a k quantized vector (VQk) of said n quantized vectors, wherein i<
k≦
n, and said k quantized vector is the closest quantized vector to said i quantized vector of a set of quantized vectors having an index k wherein i<
k≦
n; andc. an encoder for encoding a set of input vectors (V1 -VTOT) using said n quantized vectors and said threshold for said each of said n quantized vectors.
-
Specification