Vytautas Rudžionis

Recognition of Voice Commands Using Hybrid Approach

Computerized systems with voice user interfaces could save time and ease the work of healthcare practitioners. To achieve this goal voice user interface should be reliable (to recognize the commands with high enough accuracy) and properly designed (to be convenient for the user). The paper deals with hybrid approach implementation issues for the voice commands recognition. By the hybrid approach we assume the combination of several different recognition methods to achieve higher recognition accuracy. The experimental results show that most voice commands are recognized good enough but there is some set of voice commands which recognition is more complicated. In this paper the novel method is proposed for the combination of several recognition methods based on the Ripper algorithm. Experimental evaluation showed that this method allows achieve higher recognition accuracy than application of blind combination rule.

Lithuanian Digits Recognition by Using Hybrid Approach by Combining Lithuanian Google Recognizer and Some Foreign Language Recognizers

In this paper we are presenting our results obtained by experimenting with different classification methods which are suitable for creating hybrid speech recognizer. We tried to create Lithuanian digits recognizer which is capable of producing more than 95 % accuracy. Classification methods that were used are: k-Nearest neighbors (5, 11, 15, and 21), Linear Discriminant Analysis, Quadratic Discriminant Analysis, Logistic Regression, Naive Bayes, Support Vectors classifier. Experiments were taken using five foreign recognizers: English, Russian, two German and Google recognizer for Lithuanian language. Best results were received when Naive Bayes classifier was used – 97.51 %. Average accuracy of single recognizers: English – 84.9 %, Russian – 80.6 %, German I – 63 %, German II – 81.4 % and Google recognizer – 82.6 %. By using hybrid approach accuracy was increased by 12.61 % compared with best single recognizer result.

Assistive Tools for the Motor-Handicapped People Using Speech Technologies: Lithuanian Case

The paper presents analysis of the possibilities to use voice technologies for the partial integration of people with disabilities. The particular interest has been expressed to the motor-handicapped people. The special wheelchair with the voice command recognition capabilities has been designed. Evaluation of command’s recognition accuracy shows high dependency on the proper detection of the utterance boundaries. The acoustic boundaries detection algorithm has been proposed. This algorithm allowed achieve high accuracy of the detection of acoustic events boundaries such as words or phrases even in the presense of high noise. The proper detection leads to the increased accuracy of voice commands recognition and the overall satisfaction of users.

Identifying Lithuanian Native Speakers Using Voice Recognition

In this paper, we analyze speaker identification and present identification test results on Lithuanian native speakers’ database LIEPA. Two approaches for speaker acoustic modeling are examined. We start by extracting MFCC features from audio samples, then we feed this data to create speaker acoustic model with hidden Markov models (1) and with deep neural networks (2). We compare both methods by nalyzing the subset of samples from LIEPA database. This helps to achieve more than 96% identification accuracy on sample dataset.

Speaker Authentication System Based on Voice Biometrics and Speech Recognition

In this paper we are analyzing possibility to authenticate speaker by using user voice biometrics and speech recognition. Process of authentication is simple: user says his personal ID number, then system makes prediction of users’ identity and compares it to claimed ID. If these two parameters are equal system makes a positive decision. Using proposed algorithms we managed to achieve 70.45% accuracy for system, with identification module accuracy of 100% and recognition module accuracy of 70.45%.

Applications of Voice Based Informative Public Services: Lithuanian Case

The paper presents analysis of the possibilities to implement some voice based informative services in public sector of Lithuania. The paper reveals the applications in the public sector where the implementation of voice based modality may contribute significantly to the convenience and the popularity of the service. Possible solutions for the cases when Lithuanian speech will be used as the main speech in the service are evaluated both from technological and economical point of view. Several solutions how the obstacles related with the insufficient recognition accuracy of the Lithuanian voice commands may be lessened are revealed also. It is shown that implementation of properly trained foreign language speech recognizer may achieve recognition accuracy of Lithuanian voice commands which is necessary for applications in public sector.