Computerized Speech Recognition Research Articles

Speech Recognition by Integrating Hidden Markov Model Correlated with Artificial Neural Network

Now every day's speech recognition is utilized broadly in numerous packages. In software program engineering and electric constructing, speech recognition (SR) is the interpretation of verbally expressed words into textual content. it's miles otherwise referred to as "computerized speech recognition" (CSR), "pc speech reputation", or most effective "speech to text" (STT). A hid Markov model (HMM) is a measurable Markov model wherein the framework being verified is notion to be a Markov process with in mystery (shrouded) states. A HMM may be introduced as the least hard dynamic Bayesian system. Dynamic time warping (DTW) is a truly understood strategy to locate a really perfect arrangement among two given (time-subordinate) groupings underneath sure confinements instinctively; the groupings are distorted in a nonlinear manner to coordinate each other. ANN is non-immediately statistics driven self-versatile methodology. it can distinguish and research co-related examples between information dataset and evaluating target esteems. Within the wake of preparing ANN may be utilized to anticipate the end result of new unfastened facts.

A Tale of Two Cities: Evolution of Academic Physiatry in Boston and Baltimore: Part 1: The Boston Marathon

A Tale of Two Cities: Evolution of Academic Physiatry in Boston and Baltimore: Part 1: The Boston Marathon

Incidence of speech recognition errors in the emergency department

BackgroundPhysician use of computerized speech recognition (SR) technology has risen in recent years due to its ease of use and efficiency at the point of care. However, error rates between 10 and 23% have been observed, raising concern about the number of errors being entered into the permanent medical record, their impact on quality of care and medical liability that may arise. Our aim was to determine the incidence and types of SR errors introduced by this technology in the emergency department (ED). SettingLevel 1 emergency department with 42,000 visits/year in a tertiary academic teaching hospital. MethodsA random sample of 100 notes dictated by attending emergency physicians (EPs) using SR software was collected from the ED electronic health record between January and June 2012. Two board-certified EPs annotated the notes and conducted error analysis independently. An existing classification schema was adopted to classify errors into eight errors types. Critical errors deemed to potentially impact patient care were identified. ResultsThere were 128 errors in total or 1.3 errors per note, and 14.8% (n=19) errors were judged to be critical. 71% of notes contained errors, and 15% contained one or more critical errors. Annunciation errors were the highest at 53.9% (n=69), followed by deletions at 18.0% (n=23) and added words at 11.7% (n=15). Nonsense errors, homonyms and spelling errors were present in 10.9% (n=14), 4.7% (n=6), and 0.8% (n=1) of notes, respectively. There were no suffix or dictionary errors. Inter-annotator agreement was 97.8%. ConclusionsThis is the first estimate at classifying speech recognition errors in dictated emergency department notes. Speech recognition errors occur commonly with annunciation errors being the most frequent. Error rates were comparable if not lower than previous studies. 15% of errors were deemed critical, potentially leading to miscommunication that could affect patient care.

Automatic evaluation of speech impairment caused by wearing a dental appliance

In dentistry, speech evaluation is important for appropriate orofacial dysfunction rehabilitation. The speech intelligibility test is often used to assess patients’ speech, and it involves an evaluation by human listeners. However, the test has certain shortcomings, and an alternative method, without a listening procedure, is needed. The purpose of this study was to test the applicability of an automatic speech intelligibility test system using a computerized speech recognition technique. Speech of 10 normal subjects, when wearing a dental appliance, was evaluated using an automatic speech intelligibility test system that was developed using computerized speech recognition software. The results of the automatic test were referred to as the speech recognition scores. The Wilcoxon signed rank test was used to analyze differences in the results of the test between the following 2 conditions: with the palatal plate in place and with the palatal plate removed. Spearman correlation coefficients were used to evaluate whether the speech recognition score correlated with the result of conventional intelligibility test. The speech recognition score was significantly decreased when wearing the plate (z = -2.807, P = 0.0050). The automatic evaluation results positively correlated with that of conventional evaluation when wearing the appliance (r = 0.729, P = 0.017). The automatic speech testing system may be useful for evaluating speech intelligibility in denture wearers.

Application of an automatic conversation intelligibility test system using computerized speech recognition technique

Subjective speech evaluation is important in maxillofacial rehabilitation. In previous research, the automatic speech intelligibility test system using monosyllables was developed using a computerized speech recognition technique, and it was suggested that the system may effectively be used in evaluated maxillectomy patients' speech. However the system showed some disadvantages such as the low sensitiveness for the mild speech disorders. In this study, the system was remodeled employing sentences as test stimuli to overcome the shortcomings. The new system was used to evaluate speech of normal adults and maxillectomy patients. The speech of 10 normal adults and 13 maxillectomy patients was evaluated using the automatic conversation intelligibility test system that employed sentences as test stimuli. Normal subjects repeated the test for three times on different days. The patients' utterances were examined in two conditions: with and without prostheses. The results from patients with and without prostheses were compared to each other, and also compared with their corresponding conversation intelligibility obtained using a perceptual method. In normal subjects, there was no significant difference in their scores using the automatic conversation intelligibility test on three different days. The result of the automatic test was significantly improved in patients by wearing prostheses. There were significant correlations between the result of the automatic test and the corresponding conversation intelligibility both in patients with prostheses and without prostheses. It was indicated that the automatic conversation intelligibility test system using sentences as stimuli would be effectively used in evaluating maxillectomy patients' speech.

Effect of Ambient Sound Masking on the Accuracy of Computerized Speech Recognition

To evaluate the direct effect that sound masking has on computerized speech recognition (SR) accuracy in the digital reading room while eliminating the Lombard effect. This study complies with HIPAA requirements, and the institutional review board determined that approval was not required; informed consent was obtained. Ten radiologists digitally recorded 20 selected reports. Recorded reports were transcribed by using a commercial SR system at sound-masking levels 1-4, representing lowest to highest A-weighted sound pressure measurements in decibels. Dictated reports were compared with original reports to determine numbers of errors. A repeated-measures analysis of variance was used for overall comparison of mean percentage of transcription inaccuracies across sound-masking levels, and paired t tests were used to compare each sound-masking level to the baseline. Mean percentage of transcription inaccuracies at baseline was 11.6% (range, 6.3%-26.1%). Mean percentages of transcription inaccuracies at sound-masking levels 1-4 were 11.3%, 12.3%, 13.0%, and 13.6%, respectively. Mean percentages of transcription inaccuracies differed across all sound-masking levels (P < .001) but not between baseline and sound-masking level 1 (P = .313). Mean percentages of transcription inaccuracies at sound-masking levels 2-4 were each significantly higher than at baseline (P < .01 for each). Low-level sound masking provided slightly but not significantly improved SR accuracy, whereas higher levels decreased transcription accuracy. Appropriate levels of sound masking in reading room environments may decrease the negative effect of ambient noise without a deleterious effect on SR accuracy.

Speech Perception: Motoric Contributions versus the Motor Theory

Recent studies indicate that the motor cortex is involved not only in the production of speech, but also in its perception. These studies have sparked a renewed interest in gesture-based theories of speech perception.

Bilingual Computerized Speech Recognition Screening for Depression Symptoms

The Voice-Interactive Depression Assessment System (VIDAS) is a computerized speech recognition application for screening depression based on the Center for Epidemiological Studies–Depression scale in English and Spanish. Study 1 included 50 English and 47 Spanish speakers. Study 2 involved 108 English and 109 Spanish speakers. Participants completed demographic and acculturation scales, Beck Depression Inventory–II, Beck Anxiety Inventory, Composite International Diagnostic Interview–Short Form, and VIDAS (aural or visual). Both studies examined the psychometric properties and participant comfort ratings for VIDAS. Study 2 also examined psychometric sensitivity and specificity, and participant selection of digitized gender for VIDAS. Both studies found that VIDAS generally demonstrated adequate interitem reliability, convergent validity, sensitivity, and specificity. However, the aural modality displayed lower reliability and validity. Participants in both language groups rated each VIDAS modality favorably; however, the visual modality received higher positive evaluations. Participant comfort ratings of digitized gender revealed an interaction, such that the visual female and the aural male were rated more favorably.

Computerized content analysis of speech plus speech recognition in the measurement of neuropsychiatric dimensions

The Psychiatric Content Analysis and Diagnosis (PCAD) program performs automated content analysis of machine-readable transcriptions of speech samples to measure the magnitude of neuropsychiatric states and traits. Technological advances provided by computerized speech recognition may offer a possible alternative to labor-intensive manual transcription for preparation of samples for PCAD processing. To test this hypothesis, 25 digitally recorded verbal samples were transcribed both manually and by a commercially available speech recognition software package, and the transcriptions scored by PCAD. The inter-correlations between scores derived from the two different methods of transcriptions offer mixed results, with values ranging from a high of 0.920 to a low of -0.119.

A Bilingual Computerized Voice-Interactive System for Screening Depression Symptoms

ABSTRACT The Voice-Interactive Depression Assessment Sys-tem-III (VIDAS-III) is a bilingual computerized speech recognition application for screening depression. In this study, 128 English- and 128 Spanish-speakers completed a demographic interview, an acculturation scale, the BDI-II, the CIDI-Short Form, and VIDAS-III. The results suggested that VIDAS-III subscales (CES-D and DSM-IV criteria) demonstrated high inter-item reliability (.81 to .92), strong concurrent validity (.58 to .67), sensitivity (.72 to 1.0), and specificity (.40 to .72). Both language groups positively rated VIDAS-III. Male and female participants most often selected a digitized female voice to present VIDAS-III. The findings suggest that speech computerized interviewing can accurately detect depressive symptoms in English and Spanish.

Informatics in Radiology (infoRAD): radiology report entry with automatic phrase completion driven by language modeling.

Keyboard entry or correction of radiology reports by radiologists and transcriptionists remains necessary in many settings despite advances in computerized speech recognition. A report entry system that implements an automated phrase completion feature based on language modeling was developed and tested. The special text editor uses context to predict the full word or phrase being typed, updating the displayed prediction after each keystroke. At any point, pressing the tab key inserts the predicted phrase without having to type the remaining characters of the phrase. Successive words of the phrase are predicted by a trigram language model. Phrase lengths are chosen to minimize the expected number of keystrokes as predicted by the language model. Operation is highly and automatically customized to each user. The language model was trained on 36,843 general radiography reports, which were consecutively generated and contained 1.48 million words. Performance was tested on 200 randomly selected reports outside of the training set. The phrase completion technique reduced the average number of keystrokes per report from 194 to 58; the average reduction factor was 3.3 (geometric mean) (95% confidence interval, 3.2-3.5). The algorithm significantly reduced the number of keystrokes required to generate a radiography report (P <.00005).

Muscle Tension Dysphonia in Patients who use Computerized Speech recognition Systems

The use of speech recognition systems as a replacement for other types of transcription systems is increasing rapidly, partly because many people are unable to use conventional keyboards as a result of upper-extremity repetitive strain injury (RSI). However, the frequent or continuous use of such systems can cause muscle tension dysphonia in some patients. The scientific literature suggests that there is an association between upper-extremity RSI and muscle tension dysphonia. We present a retrospective case series of five patients with workplace upper-extremity RSI who developed muscle tension dysphonia soon after they began using discrete computerized speech recognition software. The diagnosis of dysphonia was based on laryngovideostroboscopy, acoustic analyses, and voice load testing. All patients had normal voice when using everyday speech, but speaking into the computer resulted in the rapid onset of aperiodicity, strain, and a decrease in fundamental frequency. In three of the five patients, laryngovideostroboscopy showed posterior glottic overapproximation, but no other abnormalities. Treatment was centered on voice therapy and avoidance of long periods of using computerized speech recognition systems. The condition of three of the five patients improved with therapy. We conclude that computer speech recognition programs can lead to the onset of muscle tension dysphonia in some patients. These patients can be successfully treated with voice therapy.

Speech recognition system which creates acoustic models by concatenating acoustic models of individual words

A computerized speech recognition system creates acoustic models of phrases by concatenating acoustic models for individual words. The system stores an acoustic word model and spelling for each of its vocabulary words. When it receives the spelling of a multi-word phrase to be treated as a new vocabulary word, it stores that multi-word spelling as the spelling of the new vocabulary word, and a new acoustic model created by concatenating the acoustic word models of previous vocabulary words whose spellings correspond to words in the multi-word spelling as the acoustic model for the new word. The system can then perform speech recognition by comparing acoustic signals against the word models of stored vocabulary words, including those representing such multi-word phrases. Preferably when a multi-word model is formed, the individual acoustic models concatenated are modified to represent the coarticulation which takes place between words spoken continuously. This can be done by representing word models as sequences of individual phonemes, individual phonemes by phoneme-in-context models, and coarticulation by modifying the phoneme-in-context models of phonemes adjacent the boundary between concatenated words models to reflect the context of the phonemes on the other side of such word boundaries. The system can be a discrete word recognizer. The multi-word phrases can be selected by a user or be obtained from other programs running on the same computer as the speech recognizer, such as from one or more commands available in another program.

Voice-activated retrieval of mammography reference images.

We undertook this project to integrate context sensitive computer-based educational and decision making aids into the film interpretation and reporting process, and to determine the clinical utility of this method as a guide for further system development. An image database of 347 digital mammography images was assembled and image features were coded. An interface was developed to a computerized speech recognition radiology reporting system which was modified to translate reported findings into database search terms. These observations were used to formulate database search strategies which not only retrieved similar cases from the image database, but also other cases that were related to the index case in different ways. The search results were organized into image sets intended to address common questions that arise during image interpretation. An evaluation of the clinical utility of this method was performed as a guide for further system development. We found that voice dictation of prototypical mammographic cases resulted in automatic retrieval of reference images. The retrieved images were organized into sets matching findings, diagnostic hypotheses, diagnosis, spectrum of findings or diagnoses, closest match to dictated case, or user specified parameters. Two mammographers graded the clinical utility of each form of system output. We concluded that case specific and problem specific image sets may be automatically generated from spoken case dictation. A potentially large number of retrieved images may be divided into subsets which anticipate common clinical problems. This automatic method of context sensitive image retrieval may provide a "continuous" form of education integrated into routine case interpretation.

Computerized speech recognition: influence of intelligibility and perceptual consistency on recognition accuracy

The effects of intelligibility and consistency on the recognition accuracy of a speaker-adaptable speech recognition system (IBM VoiceType Version 1.0) were evaluated. Six participants who had dysarthria of speech across three severity levels (i.e., mild, moderate, severe) and six age- and gender-matched peers without speech impairments participated in the study. Productions of sentences were evaluated across five assessment sessions. Recognition accuracy was significantly higher for the speakers in the control group than for the speakers with dysarthria across severity levels. High levels of intelligibility correlated significantly with high recognition accuracy scores. Perceptual rankings of speech consistency did not correlate significantly with recognition accuracy scores. Results suggest that for speaker-adaptable systems, the more intelligible a speaker, the greater his or her success with the voice recognition system. Results also suggest that perceived inconsistencies in the speech productions of dysarthric speakers may not limit their use of a speaker-adaptable speech recognition system.

Automated Screening for Depression: Toward Culturally and Linguistically Appropriate Uses of Computerized Speech Recognition

The feasibility of using computers with speech recognition capability in screening for depression was examined. Computerized and written forms of the Center for Epidemiological Studies-Depression (CES-D) scale were administered to 19 Spanish-speaking and 19 English-speaking primary care medical patients (n = 38) at a hospital-based outpatient depression clinic. A counterbalanced experimental design was employed to randomly assign the order of administration to the language groups. Psychometric analyses of the two methods suggested that the two forms were highly equivalent in both languages. There were no significant differences in the means and variances of the two forms. Ranked-order correlations and coefficient alpha estimates for interitem consistency were high. The two methods were found to be acceptable in equal proportionsfor both language groups, with a tendency to prefer the computerized method.

An exploratory study using computerized speech recognition for screening depressive symptoms

This exploratory study tested a computerized speech recognition application for screening depressive symptoms. A “talking” prototype administered the Center for Epidemiological Studies — Depression (CES-D) scale. The study compared computerized and written forms of the CES-D with an adult population. A counterbalanced design controlled for order of administration between the two methods. A sample of 68 participants completed both forms of the CES-D and a computer anxiety scale. The results suggested that there were no significant differences between the means and variances of the two CES-D methods. Equivalent forms reliability and internal consistency estimates of the two CES-D forms were very high. Correlations between the two CES-D methods and a computer anxiety scale did not differ significantly. The preference rates of the two CES-D methods were not significantly different.

Computerized speech recognition in psychological assessment: A Macintosh prototype for screening depressive symptoms

Computerized speech-recognition technology holds strong promise for psychological assessment. This paper focuses on a computerized speech recognition application for screening depressive symptomatology. A Macintosh-based prototype has been developed that uses the Center for Epidemiological Studies-Depression scale (CES-D). The prototype is a HyperCard stack interfaced with the Voice Navigator II speech recognition application. The “talking” program represents a viable depression screening tool that is fully voice operated by the respondent. A pilot study will assess the feasibility and acceptability of the computerized and written versions of the CES-D with a nonclinical adult population. A counterbalanced design will test for order effects and analyze the psychometric equivalence of the two methods. The limitations and future directions for speech-recognition applications in psychological assessment are discussed.

Pilot study on the feasibility of a computerized speech recognition charting system

The objective of this study was to determine the feasibility of developing and using a voice recognition computerized charting system to record dental clinical examination data. More specifically, the study was designed to analyze the time and error differential between the traditional examiner/recorder method (ASSISTANT) and computerized voice recognition method (VOICE). DMFS examinations were performed twice on 20 patients using the traditional ASSISTANT and the VOICE charting system. A statistically significant difference was found when comparing the mean ASSISTANT time of 2.69 min to the VOICE time of 3.72 min (P less than 0.001). No statistically significant difference was found when comparing the mean ASSISTANT recording errors of 0.1 to VOICE recording errors of 0.6 (P = 0.059). 90% of the patients indicated they felt comfortable with the dentist talking to a computer and only 5% of the sample indicated they opposed VOICE. Results from this pilot study indicate that a charting system utilizing voice recognition technology could be considered a viable alternative to traditional examiner/recorder methods of clinical charting.