ENHANCED LBG ALGORITHM USING NEURAL NETWORKS

Clustering applications cover several fields such as audio and video data compression, pattern recognition, computer vision, medical image recognition, etc. In this paper we present a new clustering algorithm called Enhanced LBG (ELBG). It belongs to the hard and K-means vector quantization groups and derives directly from the simpler LBG. The basic idea we have developed is the concept of utility of a codeword, a powerful instrument to overcome one of the main drawbacks of clustering algorithms: generally, the results achieved are not good in the case of a bad choice of the initial codebook. We will present our experimental results showing that ELBG is able to find better codebooks than previous clustering techniques and the computational complexity is virtually the same as the simpler LBG.

GENERALIZED LLOYD ALGORITHM (GLA) OR LBG

The LBG algorithm is a finite sequence of steps in which, at every step, a new quantizer, with a total distortion less or equal to the previous one, is produced

Firstly, we will describe the optimization step. It will simplify the LBG explanation. In fact, several concepts necessary to describe this step are useful for the initialization phase, too. In the following we will use these symbols:

• m: iteration number;

• Ym: mth codebook;

• Dm: MQE calculated at the mth iteration.