Lynn D. Wilcox

Wordspotting for voice editing and indexing

A technique for wordspotting based on hidden Markov models (HMMs). The technique allows a speaker to specify keywords dynamically and to train the associated HMMs via a single repetition of a keyword. Non-keyword speech is modeled using an HMM trained from a prerecorded sample of continuous speech. The wordspotter is intended for interactive applications, such as the editing of voice mail or mixed-media documents, and for keyword indexing in single-speaker audio or video recordings.

Character and phoneme recognition based on probability clustering

Prior to character or phoneme recognition, a classifier provides a respective probability list for each of a sequence of sample characters or phonemes, each probability list indicating the respective samples probability for each character or phoneme type. These probability lists are clustered in character or phoneme probability space, in which each dimension corresponds to the probability that a character or phoneme candidate is an instance of a specific character or phoneme type. For each resulting cluster, data is stored indicating its cluster ID and a probability list indicating the probability of each type at the clusters center. Then, during recognition, a probability cluster identifier compares the probability list for each candidate with the probability list for each cluster to find the nearest cluster. The cluster identifier then provides the nearest clusters cluster ID to a constraint satisfier that attempts to recognize the candidate based on rules, patterns, or a combination of rules and patterns. If necessary, the constraint satisfier uses the cluster ID to retrieve the stored probability list of the cluster to assist it in recognition.

Method of speaker clustering for unknown speakers in conversational audio data

A method for clustering speaker data from a plurality of unknown speakers. The method includes steps of providing a portion of audio data containing speech from at least all the speakers in the audio data and dividing the portion into data clusters. A pairwise distance between each pair of clusters is computed, the pairwise distance being based on a likelihood that two clusters were created by the same speaker, the likelihood measurement being biased by the prior probability of speaker changes. The two clusters with a minimum pairwise distance are combined into a new cluster and speakers models are trained for each of the remaining clusters including the new cluster. The likelihood that two clusters were created by the same speaker may be biased by a Markov duration model based on speaker changes over the length of the initial data clusters.

Detection and location of multicharacter sequences in lines of imaged text

A system for detecting and locating user-specified search strings, or phrases, in lines of imaged text is described. The phrases may be single words or multiple words, and may contain a partially specified word. The imaged text can be composed of a number of different fonts and graphics. Textlines in a deskewed image are hypothesized using multiresolution morphology. For each textline, the baseline, topline and x-height are identified by simple statistical methods and then used to normalize each textline bounding box. Columns of pixels in the resulting bounding box serve as feature vectors. One hidden Markov model is created for each user-specified phrase and another represents all text and graphics other than the user-specified phrases. Phrases are identified using Viterbi decoding on a spotting network created from the models. The operating point of the system can be varied to trade off the percentage of words correctly spotted and the percentage of false alarms. Results are given using a subset of the UW English Document Image Database I.

Audio indexing using speaker identification

In this paper, a technique for audio indexing based on speaker identification is proposed. When speakers are known a priori, a speaker index can be created in real time using the Viterbi algorithm to segment the audio into intervals from a single talker. Segmentation is performed using a hidden Markov model network consisting of interconnected speaker sub- networks. Speaker training data is used to initiate sub-networks for each speaker. Sub- networks can also be used to model silence, or non-speech sounds such as musical theme. When no prior knowledge of the speakers is available, unsupervised segmentation is performed using a non-real time iterative algorithm. The speaker sub-networks are first initialized, and segmentation is performed by iteratively generating a segmentation using the Viterbi algorithm, and retraining the sub-networks based on the results of the segmentation. Since the accuracy of the speaker segmentation depends on how well the speaker sub-networks are initiated, agglomerative clustering is used to approximately segment the audio according to speaker for initialization of the speaker sub-networks. The distance measure for the agglomerative clustering is a likelihood ratio in which speed segments are characterized by Gaussian distributions. The distance between merged segments is recomputed at each stage of the clustering, and a duration model is used to bias the likelihood ratio. Segmentation accuracy using agglomerative clustering initialization matches accuracy using initialization with speaker labeled data.

Introduction to the human factors and usability issues minitrack

Usability Engineering is the newest buzz word of the millennium. Today’s customer is well aware of his requirements and is unwilling to compromise. He wants value for his hard-earned money. Whether he purchases a mobile phone, a microwave oven, or a washing machine – the focus has now shifted from features offered, to the ease and convenience of operation, and how fast the gadget can be mastered – i.e. the focus is now on the “USER INTERFACE”.The development and use digital documents must address the context in which these digital documents will be used and what roles they will play at a group level, an institutional level, or an inter-organisational level. For a given task, the usefulness of a tool depends on many factors. There is clearly a need for a forum to discuss human factors and empirical studies of digital document systems.This minitrack aims to provide an inter-disciplinary forum for researchers and practitioners to communicate their work in relation to user-centred design and evaluation issues in the larger context of digital documents. The focus is on the role of human factors in information-intensive environments in which digital documents have an integral role. This minitrack has grouped articles under two broad themes: Usability and Beyond, and Design in Context.