Ghinwa F. Choueiter

An empirical study of automatic accent classification

This paper extends language identification (LID) techniques to a large scale accent classification task: 23-way classification of foreign-accented English. We find that a purely acoustic approach based on a combination of heteroscedastic linear discriminant analysis (HLDA) and maximum mutual information (MMI) training is very effective. In contrast to LID tasks, methods based on parallel language models prove much less effective. We focus on the Oregon Graduate Institute Foreign-Accented English dataset, and obtain a detection rate of 32%, which to our knowledge is the best reported result for 23-way accent classification.

An Implementation of Rational Wavelets and Filter Design for Phonetic Classification

Although wavelet analysis has been proposed for speech processing as an alternative to Fourier analysis, most approaches make use of off-the-shelf wavelets and dyadic tree-structured filter banks. In this paper, we extend previous wavelet-based frameworks in two ways. First, we increase the flexibility in wavelet selection by taking advantage of the relationship between wavelets and filter banks and by designing new wavelets using filter design methods. We adopt two filter design techniques that we refer to as filter matching and attenuation minimization. Second, we improve the flexibility in frequency partitioning by implementing rational as well as dyadic filter banks. Rational filter banks naturally incorporate the critical-band effect in the human auditory system. To test our extensions, we implement an energy-based measurement which we also compare in performance to the mel-frequency cepstral coefficients (MFCCs) in a phonetic classification task. We show that the designed wavelets outperform off-the-shelf wavelets as well as an MFCC baseline

Automatic lexical pronunciations generation and update

Most automatic speech recognizers use a dictionary that maps words to one or more canonical pronunciations. Such entries are typically hand-written by lexical experts. In this research, we investigate a new approach for automatically generating lexical pronunciations using a linguistically motivated subword model, and refining the pronunciations with spoken examples. The approach is evaluated on an isolated word recognition task with a 2 k lexicon of restaurant and street names. A letter-to-sound model is first used to generate seed baseforms for the lexicon. Then spoken utterances of words in the lexicon are presented to a subword recognizer and the top hypotheses are used to update the lexical base-forms. The spelling of each word is also used to constrain the subword search space and generate spelling-constrained baseforms. The results obtained are quite encouraging and indicate that our approach can be successfully used to learn valid pronunciations of new words.

A turbo-style algorithm for lexical baseforms estimation

In this research, an iterative and unsupervised Turbo-style algorithm is presented and implemented for the task of automatic lexical acquisition. The algorithm makes use of spoken examples of both spellings and words and fuses information from letter and subword recognizers to boost the overall lexical learning performance. The algorithm is tested on a challenging lexicon of restaurant and street names and evaluated in terms of spelling accuracy and letter error rate. Absolute improvements of 7.2% and 3% (15.5% relative improvement) are obtained in the spelling accuracy and the letter error rate respectively following only 2 iterations of the algorithm.

A novel parallel architecture for local histogram equalization

Local histogram equalization is an image enhancement algorithm that has found wide application in the pre-processing stage of areas such as computer vision, pattern recognition and medical imaging. The computationally intensive nature of the procedure, however, is a main limitation when real time interactive applications are in question. This work explores the possibility of performing parallel local histogram equalization, using an array of special purpose elementary processors, through an HDL implementation that targets FPGA or ASIC platforms. A novel parallelization scheme is presented and the corresponding architecture is derived. The algorithm is reduced to pixel-level operations. Processing elements are assigned image blocks, to maintain a reasonable performance-cost ratio. To further simplify both processor and memory organizations, a bit-serial access scheme is used. A brief performance assessment is provided to illustrate and quantify the merit of the approach.