Fast Algorithm and Efficient Implementation of GMM-Based Pattern Classifiers

Authors:
Hidenori Watanabe;Shogo Muramatsu
Affiliations:
Department of Electrical and Electronic Engineering, Faculty of Engineering, Niigata University, Niigata, Japan 950-2181;Department of Electrical and Electronic Engineering, Faculty of Engineering, Niigata University, Niigata, Japan 950-2181
Venue:
Journal of Signal Processing Systems
Year:
2011

Citing 9
Cited 0

Statistical Pattern Recognition: A Review

IEEE Transactions on Pattern Analysis and Machine Intelligence
Bayesian model search for mixture models based on optimizing variational bounds

Neural Networks
Skin Segmentation Using Color Pixel Classification: Analysis and Comparison

IEEE Transactions on Pattern Analysis and Machine Intelligence
Pattern Recognition and Machine Learning (Information Science and Statistics)

Pattern Recognition and Machine Learning (Information Science and Statistics)
An efficient digital VLSI implementation of Gaussian mixture models-based classifier

IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Numerical Recipes 3rd Edition: The Art of Scientific Computing

Numerical Recipes 3rd Edition: The Art of Scientific Computing
Digital Signal Processing with Field Programmable Gate Arrays

Digital Signal Processing with Field Programmable Gate Arrays
A Study of Variable-Parameter Gaussian Mixture Hidden Markov Modeling for Noisy Speech Recognition

IEEE Transactions on Audio, Speech, and Language Processing
Variational learning for Gaussian mixture models

IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper proposes a fast decision algorithm in pattern classification based on Gaussian mixture models (GMM). Statistical pattern classification problems often meet a situation that comparison between probabilities is obvious and involve redundant computations. When GMM is adopted for the probability model, the exponential function should be evaluated. This work firstly reduces the exponential computations to simple and rough interval calculations. The exponential function is realized by scaling and multiplication with powers of two so that the decision is efficiently realized. For finer decision, a refinement process is also proposed. In order to verify the significance, experimental results on TI DM6437 EVM board and TED TB-3S-3400DSP-IMG board are shown through the application to a color extraction problem. It is verified that the classification was almost completed without any refinement process and the refinement process can proceed the residual decisions.