Keun-Chang Kwak

Automated Speaker Recognition for Home Service Robots Using Genetic Algorithm and Dempster–Shafer Fusion Technique

An automated speaker recognition system for home service robots is proposed in this paper. In an uncontrolled environment, a speech classifier should be adaptive to different users and robust to noisy environments. It is usually observed that specific features and classifiers are more appropriate to parts of the problem domain than others; therefore, we propose a self-optimizing approach in which multiple feature extraction and classification techniques are simultaneously considered. The system uses a genetic algorithm to simultaneously select features and classifier, and the results from multiple classifiers are then combined using the Dempster-Shafer theory. The set of feature extractors used here includes linear-prediction coefficients, linear-prediction cepstral coefficients, mel-frequency cepstral coefficients, and bark-frequency cepstral coefficients, and the set of classifiers includes the Gaussian mixture model, support vector machines, C4.5 decision tree, k nearest neighbors, and multilayer perceptron neural network. The WEVER-R2 home service robot is used in a typical Korean home environment to collect speech signals for evaluating the performance of the proposed system for gender and age classification. Classification results show that the performance of the proposed method consistently outperforms the individual classifiers.

AR-KLT based Hand Tracking

This paper proposes a novel real-time robust hand tracking algorithm, integrating multi-cues, and a limbs degree of freedom. For this purpose, we construct a limb model and maintain the model obtained from KLT-AR methods with respect to second-order auto-regression model and Kanade-Lucas-Tomasi (KLT) features, respectively. Furthermore, this method provides directivity of a target, enabling us to predict the next motion. Thus, we can develop a method of hand tracking for gesture and behavior recognition techniques frequently used in conjunction with human-robot interaction (HRI) components. The experimental results show that the proposed method yields a good performance in the intelligent service robots, so called Wever developed in ETRI

Classification of K-Pop Dance Movements Based on Skeleton Information Obtained by a Kinect Sensor

This paper suggests a method of classifying Korean pop (K-pop) dances based on human skeletal motion data obtained from a Kinect sensor in a motion-capture studio environment. In order to accomplish this, we construct a K-pop dance database with a total of 800 dance-movement data points including 200 dance types produced by four professional dancers, from skeletal joint data obtained by a Kinect sensor. Our classification of movements consists of three main steps. First, we obtain six core angles representing important motion features from 25 markers in each frame. These angles are concatenated with feature vectors for all of the frames of each point dance. Then, a dimensionality reduction is performed with a combination of principal component analysis and Fisher’s linear discriminant analysis, which is called fisherdance. Finally, we design an efficient Rectified Linear Unit (ReLU)-based Extreme Learning Machine Classifier (ELMC) with an input layer composed of these feature vectors transformed by fisherdance. In contrast to conventional neural networks, the presented classifier achieves a rapid processing time without implementing weight learning. The results of experiments conducted on the constructed K-pop dance database reveal that the proposed method demonstrates a better classification performance than those of conventional methods such as KNN (K-Nearest Neighbor), SVM (Support Vector Machine), and ELM alone.

Reliable Speaker Identification Using Multiple Microphones in Ubiquitous Robot Companion Environment

This paper presents a text-independent speaker identification system using multiple microphones on the robot, which is intended for use in human-robot interaction. For the purpose of the best possible classification rate in speaker identification, the individual identification results obtained from multiple microphones on the robot are combined by various combination schemes. The performance improvement has been achieved. Our ultimate goal is to enhance human-robot interaction by improving the recognition performance of speaker identification with multiple microphones on the robot side in adverse distant-talking environments. Various combination schemes obtained high classification accuracy in the ubiquitous robot companion (URC) environment, where the robot is connected to a server through extremely high broadband penetration rate. In conclusion, our speaker identification system can provide human-robot interaction with a reliable basic interface with high classification accuracy.

Enhanced Speech Recognition with Blind Equalization For Robot "WEVER-R2"

Ambient noise caused by room acoustics usually degrades quality and reliability of speech recognition for the service robot application. In this paper, we propose enhancing the robustness of speech recognition for home service robots by combining blind equalization with classification. In particular, the linear predictive code (LPC) is used for feature extraction. The constant modulus algorithm (CMA) is combined with the radial basis function (RBF) neural network to form a robust classifier with automatic equalization capability to reduce the room service effect. Using real speech data collected by WEVER-R2 acoustic robot, it is shown that the proposed method can increase the speech recognition rate from 74.46% to 87.23% for age recognition and from 80.95% to 95.09% for gender recognition.

An incremental adaptive neuro-fuzzy networks

In this paper, we propose a method for constructing an incremental adaptive neuro-fuzzy network (IANFN). In contrast to typical rule-based systems, the underlying principle is to consider a two-step development of adaptive neuro-fuzzy network (ANFN). First, we build a standard linear regression (LR) model which could be treated as a preliminary design capturing the linear part of the data. Next, all modeling discrepancies are compensated by a collection of rules that become attached to the regions of the input space in which the error becomes localized. The incremental network is constructed by building a collection of information granules through some specialized fuzzy clustering, called context-based fuzzy c-means (CFCM) that is guided by the distribution of error of the linear part of its development. The experimental results reveal that the proposed incremental network shows a good approximation and generalization capability in comparison with the general method.

Integrating Evidences of Independently Developed Face and Speaker Recognition Systems by Using Discrete Probability Density Function

User recognition is one of the most fundamental functionalities for intelligent service robots. However, in robot applications, the conditions are far severer compared to the traditional biometric security systems. The robots should be able to recognize users non-intrusively, which confines the available biometric features to face and voice. Also, the robots are expected to recognize users from relatively afar, which inevitably deteriorates the accuracy of each recognition module. In this paper, we tried to improve the overall accuracy by integrating the evidences issued by independently developed face and speaker recognition modules. Each recognition module exhibits different statistical characteristics in representing its confidence of the recognition. Therefore, it is essential to transform the evidences to a normalized form to integrate the results. This paper introduces a novel approach to integrate mutually independent multiple evidences to achieve an improved performance. Typical approach to this problem is to model the statistical characteristics of the evidences by well-known parametric form such as Gaussian. Using Mahalanobis distance is a good example. However, the characteristics of the evidences often do not fit into the parametric models, which results in performance degradation. To overcome this problem, we adopted a discrete PDF that can model the statistical characteristics as it is. To confirm the validity of the proposed method, we used a multi-modal database that consists of 10 registered users and 550 probe data. Each probe data contains face image and voice signal. Face and speaker recognition modules are applied to generate respective evidences. The experiment showed an improvement of 11.27% in accuracy compared to the individual recognizers, which is 2.72% better than the traditional Mahalanobis distance approach.

Appearance-based Face Recognition from robot camera images with Illumination and Distance Variations

This paper is concerned with the appearance-based face recognition from robot camera images with illumination and distance variations. The approaches used in this paper consist of eigenface, fisherface, and icaface, which are the most representative recognition techniques frequently used in conjunction with face recognition. These approaches are based on a popular unsupervised and supervised statistical technique that supports finding useful image representations, respectively. Thus we focus on the performance comparison from robot camera images with unwanted variations. The comprehensive experiments are completed for two face databases with illumination and distance variations. A comparative analysis demonstrates that ICA comes with improved classification rates when compared with other approaches such as eigenface and fisherface

Bimanual hand tracking

This paper proposes a novel real-time hand tracking algorithm in the presence of occlusion. For this purpose, we construct a limb model and maintain the model obtained from ARKLT methods with respect to second-order auto-regression model and Kanade-Lucas-Tomasi(KLT) features, respectively. Furthermore, this method do not require to categorize types of superimposed hand motion based on directivity obtained by the slope’s direction of KLT regression. Thus, we can develop a method of hand tracking for gesture and activity recognition techniques frequently used in conjunction with Human-Robot Interaction (HRI) components.

Speech-based Human-Robot Interaction Components for URC Intelligent Service Robots

This video is concerned with a scenario for developing speech-based Human-Robot Interaction (sHRI) components for Ubiquitous Robot Companion (URC) Intelligent Service Robots as one of the next-generation growth engine industries in Korea. Here the URC means that it will provide the necessary services at any time and place to meet the users requirements. Thus, it combines the network function with the current concept of a robot in order to enhance mobility and human interface. The main characteristics of this video are to combine of text-independent speaker recognition and Korean-based spontaneous speech recognition from two speakers in scenario. Especially each speaker communicates with service robot through spontaneous speech recognition with continuous words to provide useful information such as daily life schedule and TV program suitable to the speaker recognized. On the basis of these components, consumers will be able to utilize various speech-based services of the robot.