Probabilistic scoring using decision trees for fast and scalable speaker recognition

Abstract

To cite this version:Gilles Gonon, Frédéric Bimbot, Rémi Gribonval. Probabilistic scoring using decision trees for fast and scalable speaker recognition. Speech Communication, Elsevier : North-Holland, 2009, 51 (11) AbstractIn the context of fast and low cost speaker recognition, this article investigates several techniques based on decision trees. A new approach is introduced where the trees are used to estimate a score function rather than returning a decision among classes. This technique is developed to approximate the GMM log-likelihood ratio (LLR) score function. On top of this approach, different solutions are derived to improve the accuracy of the proposed trees. The first one studies the quantization of the LLR function to create classification trees on the LLR values. The second one makes use of knowledge on the GMM distribution of the acoustic features in order to build oblique trees. A third extension consists in using a low-complexity score function in each of the tree leaves. Series of comparative experiments are performed on the NIST 2005 speaker recognition evaluation data in order to evaluate the impact of the proposed improvements in terms of efficiency, execution time and algorithmic complexity. Considering a baseline system with an equal error rate (EER) of 9.6% on the NIST 2005 evaluation, the best tree-based configuration achieves an EER of 12.9%, with a computational cost adapted to embedded devices and an execution time suitable for real-time speaker identification.

Biometric based personal authentication using fuzzy binary decision tree

“…A DT based face recognition using local binary patterns (LBP) is discussed in [4]. A fast and low cost speaker verification using DT is investigated in [5]. In this, DT is used to estimate a score function based on Gaussian Mixture Model (GMM) log-likelihood ratio.…”

Differential evolution based score level fusion for multi-modal biometric systems

“…BioID system, which was developed by Frischholz et al [9] associates different decision strategies for different biometric modalities but lacks from the automatic generation of adaptive strategies for desired performance. Some attempts in decision level fusion have also been made using k-NN (k-Nearest Neighbor) [10] [11], neural networks(NN) [12] [13], decisiontree(DT) [14] [15] [16] [17] , and support vector machine(SVM) [18] [19]. While considering the scaled independent contribution of input data to the output target vector, support vector machine (SVM), logistic regression and nearest neighbor methods produce satisfactory results.…”