• The KIPS Transactions:PartB
• /
• v.8B no.6
• /
• pp.699-704
• /
• 2001

ABSTRACT In this paper, a new hybrid modular recurrent TDNN (time-delay neural network)-HMM (hidden Markov model) architecture for speech recognition has been studied. In TDNN, the recognition rate could be increased if the signal window is extended. To obtain this effect in the neural network, a high-level memory generated through a feedback within the first hidden layer of the neural network unit has been used. To increase the ability to deal with the temporal structure of phonemic features, the input layer of the network has been divided into multiple states in time sequence and has feature detector for each states. To expand the network from small recognition task to the full speech recognition system, modular construction method has been also used. Furthermore, the neural network and HMM are integrated by feeding output vectors from the neural network to HMM, and a new parameter smoothing method which can be applied to this hybrid system has been suggested.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.06a
• /
• pp.758-761
• /
• 2007

Digital signal processing technique is applied in wide fields such as speech processing, image processing and spectrum analysis. Therefore, in order to do frequency selective operation digital filter is used in stead of analog filter and sharp filter characteristics can be implemented. Since finite impulse response (FIR) digital filter as nonrecursive type represents linear phase response characteristics and is always stable and is used in fields regarding wave information importantly such as data transmission. And due to frequency characteristics, in order to remove the Gibbs phenomenon generating around a discontinuous point, filter is designed through window function method. Therefore, in this paper to improve performance of FIR digital filter, a modified window function was applied. And the proposed method was compared with conventional methods using peak side-lobe and transition properties in simulations.

• Journal of the Institute of Convergence Signal Processing
• /
• v.2 no.4
• /
• pp.9-15
• /
• 2001

In this paper we investigate context dependent acoustic models to improve the performance of the Korean speech recognition . The algorithm are using the Korean phonological rules and decision tree, By Successive State Splitting(SSS) algorithm the Hidden Merkov Netwwork(HM-Net) which is an efficient representation of phoneme-context-dependent HMMs, can be generated automatically SSS is powerful technique to design topologies of tied-state HMMs but it doesn't treat unknown contexts in the training phoneme contexts environment adequately In addition it has some problem in the procedure of the contextual domain. In this paper we adopt a new state-clustering algorithm of SSS, called Phonetic Decision Tree-based SSS (PDT-SSS) which includes contexts splits based on the Korean phonological rules. This method combines advantages of both the decision tree clustering and SSS, and can generated highly accurate HM-Net that can express any contexts To verify the effectiveness of the adopted methods. the experiments are carried out using KLE 452 word database and YNU 200 sentence database. Through the Korean phoneme word and sentence recognition experiments. we proved that the new state-clustering algorithm produce better phoneme, word and continuous speech recognition accuracy than the conventional HMMs.

• Journal of rehabilitation welfare engineering & assistive technology
• /
• v.3 no.1
• /
• pp.27-34
• /
• 2009

Hearing impaired person's hearing loss has personally various shape, so existing symmetrical auditory filter of frequency band method wasn't properly simulated the hearing impaired person's various hearing loss shape. The shapes of auditory filter are asymmetrical different with each center frequency and each input level. Hearing impaired person which has hearing loss was differently changed with that of normal hearing people and it has different value for speech of quality through auditory filter. In this study, the asymmetrical auditory filter was simulated and then some tests to estimate the filter's performance objectively were performed. The experiment as simulated auditory filter's performance evaluation method used perceptual evaluation of speech quality (PESQ) and log likelihood ratio (LLR) for speech through auditory filter. In the test, processed speech was evaluated objective speech quality and distortion using PESQ and LLR value. When hearing loss processed, PESQ and LLR value have big difference between symmetrical and asymmetrical auditory filter. It means that the difference of the shape auditory filter may affect to speech quality. Especially, when hearing loss existed, auditory filter changing according to asymmetrical shape for each center frequency affected to perceive speech quality of the hearing impaired.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.6
• /
• pp.572-577
• /
• 2009

Large corpus-based concatenating Text-to-Speech (TTS) systems can generate natural synthetic speech without additional signal processing. To prune the redundant speech segments in a large speech segment DB, we can utilize a decision-tree based triphone clustering algorithm widely used in speech recognition area. But, the conventional methods have problems in representing the acoustic transitional characteristics of the phones and in applying context questions with hierarchic priority. In this paper, we propose a new clustering algorithm to downsize the speech DB. Firstly, three 13th order MFCC vectors from first, medial, and final frame of a phone are combined into a 39 dimensional vector to represent the transitional characteristics of a phone. And then the hierarchically grouped three question sets are used to construct the triphone trees. For the performance test, we used DTW algorithm to calculate the acoustic similarity between the target triphone and the triphone from the tree search result. Experimental results show that the proposed method can reduce the size of speech DB by 23% and select better phones with higher acoustic similarity. Therefore the proposed method can be applied to make a small sized TTS.

• Proceedings of the KOSOMBE Conference
• /
• v.1993 no.05
• /
• pp.134-135
• /
• 1993

The purpose of this research is development of EGG system for quantitative assessment of laryngeal function using speech and electroglotto-graphic data. The designed EGG system is 4-electrodes system which excitation current source is supplied from 1st to 4th electrode. The output signal.: from 2nd and 3rd electrodes, which are motivated by frequency of excitation current source, are air-pressure waveforms from vocal folds. After demodulation process, we obtain pitch signals of the modulated waveforms by excitation current source through differentiator which cuts off frequency below 0.1Hz. Software processing methods were used as conventional pitch extraction methods, but the proposed system is designed to analog hardware in order to eliminate interferences from low formant frequency of speech. We will construct the discriminating database between pathological subjects and control groups on each case. Using the proposed 3 channel EGG system and LMS algorithm, it will be detected that the distinctive characteristics of laryngeal function of voiced region and other regions by EGG signals and LPC spectra.

• Proceedings of the IEEK Conference
• /
• summer
• /
• pp.92-96
• /
• 2004

The researches of the emotions are currently great interest in speech processing as well as in human-machine interaction domain. In the recent years, more and more of researches relating to emotion synthesis or emotion recognition are developed for the different purposes. Each approach uses its methods and its various parameters measured on the speech signal. In this paper, we proposed using a short-time parameter: MFCC coefficients (Mel­Frequency Cepstrum Coefficients) and a simple but efficient classifying method: Vector Quantification (VQ) for speaker-dependent emotion recognition. Many other features: energy, pitch, zero crossing, phonetic rate, LPC... and their derivatives are also tested and combined with MFCC coefficients in order to find the best combination. The other models: GMM and HMM (Discrete and Continuous Hidden Markov Model) are studied as well in the hope that the usage of continuous distribution and the temporal behaviour of this set of features will improve the quality of emotion recognition. The maximum accuracy recognizing five different emotions exceeds $88\%$ by using only MFCC coefficients with VQ model. This is a simple but efficient approach, the result is even much better than those obtained with the same database in human evaluation by listening and judging without returning permission nor comparison between sentences [8]; And this result is positively comparable with the other approaches.

• Proceedings of the KIEE Conference
• /
• 2007.07a
• /
• pp.1777-1778
• /
• 2007

In this paper, We design and implement a humanoid robot, With Educational purpose, which can collaborate and communicate with human. We present an affective human-robot communication system for a humanoid robot, D2, which we designed to communicate with a human through dialogue. D2 communicates with humans by understanding and expressing emotion using facial expressions, voice, gestures and posture. Interaction between a human and a robot is made possible through our affective communication framework. The framework enables a robot to catch the emotional status of the user and to respond appropriately. As a result, the robot can engage in a natural dialogue with a human. According to the aim to be interacted with a human for voice, gestures and posture, the developed Educational humanoid robot consists of upper body, two arms, wheeled mobile platform and control hardware including vision and speech capability and various control boards such as motion control boards, signal processing board proceeding several types of sensors. Using the Educational humanoid robot D2, we have presented the successful demonstrations which consist of manipulation task with two arms, tracking objects using the vision system, and communication with human by the emotional interface, the synthesized speeches, and the recognition of speech commands.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.2
• /
• pp.192-198
• /
• 2022

As an effective weapon system for anti-submarine warfare, DIrectional Frequency Analysis and Recording (DIFAR) sonobuoy detects underwater targets via beamforming with three channels composed of an omni-direcitonal and two directional channels. However, ambient noise degrades the detection performance of DIFAR sonobouy in specific direction (0°, 90°, 180°, 270°). Thus, an ambient noise redcution technique is proposed for performance improvement of acoustic target detection of DIFAR sonobuoy. The proposed method is based on OTA (Order Truncate Average), which is widely used in sonar signal processing area, for ambient noise estimation and Wiener filtering, which is widely used in speech signal processing area, for noise reduction. For evaluation, we compare mean square errors of target bearing estmation results of conventional and proposed methods and we confirmed that the proposed method is effective under 0 dB signal-to-noise ratio.

• Journal of Satellite, Information and Communications
• /
• v.12 no.1
• /
• pp.101-106
• /
• 2017

In this paper, we present Emergency exit signs are installed to provide escape routes or ways in buildings like shopping malls, hospitals, industry, and government complex, etc. and various other places for safety purpose to aid people to escape easily during emergency situations. In case of an emergency situation like smoke, fire, bad lightings and crowded stamped condition at emergency situations, it's difficult for people to recognize the emergency exit signs and emergency doors to exit from the emergency building areas. This paper propose an automatic emergency exit sing recognition to find exit direction using a smart device. The proposed approach aims to develop an computer vision based smart phone application to detect emergency exit signs using the smart device camera and guide the direction to escape in the visible and audible output format. In this research, a CAMShift object tracking approach is used to detect the emergency exit sign and the direction information extracted using template matching method. The direction information of the exit sign is stored in a text format and then using text-to-speech the text synthesized to audible acoustic signal. The synthesized acoustic signal render on smart device speaker as an escape guide information to the user. This research result is analyzed and concluded from the views of visual elements selecting, EXIT appearance design and EXIT's placement in the building, which is very valuable and can be commonly referred in wayfinder system.